Skip to content

Instantly share code, notes, and snippets.

@kevindavenport

kevindavenport/lending_club_python.ipynb

Last active August 4, 2021 12:47
Show Gist options
  • Save kevindavenport/998c9ec12706ab9a2ef0 to your computer and use it in GitHub Desktop.
Save kevindavenport/998c9ec12706ab9a2ef0 to your computer and use it in GitHub Desktop.
Download ZIP
Display the source blob
Display the rendered blob
Raw
lending_club_python.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll download the 2013-2014 data and uncompress it from inside our notebook by invoking the command line (I didn't feel like installing wget on OSX but at least we have curl:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"%%bash\n",
"curl https://resources.lendingclub.com/LoanStats3c_securev.csv.zip | tar -xf-"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's also take a quick look at the data via bash (file size, head, line count, column count)."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"!du -h LoanStats3c_securev.csv"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"!head -n 2 LoanStats3c_securev.csv"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Examining the data we see that most of feature names are intuitive. We can get the specifics from the provided data dictionary at https://www.lendingclub.com/info/download-data.action"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"!wc -l < LoanStats3c_securev.csv"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"!head -2 LoanStats3c_securev.csv | sed 's/[^,]//g' | wc -c"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It looks we have 235,630 entries and 57 attributes. At a glance some of the more notable atributes are:\n",
"\n",
"1. Amount Requested\n",
"2. Amount Funded by Investors*\n",
"3. Interest Rate\n",
"4. Term (Loan Length)\n",
"5. Purpose of Loan\n",
"6. Debt/Income Ratio **\n",
"7. State\n",
"8. Rent or Own Home\n",
"9. Annual Income\n",
"10. 30+ days past-due incidences of delinquency in the borrower's credit file for the past 2 years\n",
"10. FICO Low\n",
"11. FICO High\n",
"12. Last FICO Low\n",
"13. Last FiCO High\n",
"12. Credit Lines Open\n",
"13. Revolving Balance\n",
"14. Inquiries in Last 6 Months\n",
"15. Length of Employment\n",
"\n",
"The last vs current FICO scores did not exist the last time we looked at the data, I'm thinking they added this to asses directionality or momentum, although this few of samples is suspicious. Interestingly enough when looking through the attributes provided for declined loans I found a \"score\" attribute which was desribed as such: \n",
"\n",
">*For applications prior to November 5, 2013 the risk score is the borrower's FICO score. For applications after November 5, 2013 the risk score is the borrower's [Vantage](https://en.wikipedia.org/wiki/VantageScore) score.*\n",
"\n",
"I suspect they are actually using both.\n",
"\n",
"*Lending club states that the amount funded by investors has no affect on the final interest rate assigned to a loan.\n",
"\n",
"** Debt/Income Ratio ratio takes all of your monthly liabilities and divides the total by your gross monthly income.\n",
"\n",
"# 1. Our Data\n",
"Let's use Pandas to manipulate our data and find an excuse to use the new pipe operators http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.pipe.html\n",
"\n",
"Our data is 126mb on disk, 102.2 in memory, not the greatest reduction in side I've seen, but we have a lot of inefficient `object` data types in the df to start."
]
},
{
"cell_type": "code",
"execution_count": 98,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"%matplotlib inline "
]
},
{
"cell_type": "code",
"execution_count": 99,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 235629 entries, 0 to 235628\n",
"Data columns (total 56 columns):\n",
"id 235629 non-null int64\n",
"member_id 235629 non-null int64\n",
"loan_amnt 235629 non-null int64\n",
"funded_amnt 235629 non-null int64\n",
"funded_amnt_inv 235629 non-null int64\n",
"term 235629 non-null object\n",
"int_rate 235629 non-null object\n",
"installment 235629 non-null float64\n",
"grade 235629 non-null object\n",
"sub_grade 235629 non-null object\n",
"emp_title 222393 non-null object\n",
"emp_length 235629 non-null object\n",
"home_ownership 235629 non-null object\n",
"annual_inc 235629 non-null float64\n",
"verification_status 235629 non-null object\n",
"issue_d 235629 non-null object\n",
"loan_status 235629 non-null object\n",
"pymnt_plan 235629 non-null object\n",
"url 235629 non-null object\n",
"desc 15279 non-null object\n",
"purpose 235629 non-null object\n",
"title 235629 non-null object\n",
"zip_code 235629 non-null object\n",
"addr_state 235629 non-null object\n",
"dti 235629 non-null float64\n",
"delinq_2yrs 235629 non-null int64\n",
"earliest_cr_line 235629 non-null object\n",
"fico_range_low 235629 non-null int64\n",
"fico_range_high 235629 non-null int64\n",
"inq_last_6mths 235629 non-null int64\n",
"mths_since_last_delinq 119748 non-null float64\n",
"mths_since_last_record 41524 non-null float64\n",
"open_acc 235629 non-null int64\n",
"pub_rec 235629 non-null int64\n",
"revol_bal 235629 non-null int64\n",
"revol_util 235504 non-null object\n",
"total_acc 235629 non-null int64\n",
"initial_list_status 235629 non-null object\n",
"out_prncp 235629 non-null float64\n",
"out_prncp_inv 235629 non-null float64\n",
"total_pymnt 235629 non-null float64\n",
"total_pymnt_inv 235629 non-null float64\n",
"total_rec_prncp 235629 non-null float64\n",
"total_rec_int 235629 non-null float64\n",
"total_rec_late_fee 235629 non-null float64\n",
"recoveries 235629 non-null float64\n",
"collection_recovery_fee 235629 non-null float64\n",
"last_pymnt_d 235486 non-null object\n",
"last_pymnt_amnt 235629 non-null float64\n",
"next_pymnt_d 200795 non-null object\n",
"last_credit_pull_d 235602 non-null object\n",
"last_fico_range_high 235629 non-null int64\n",
"last_fico_range_low 235629 non-null int64\n",
"collections_12_mths_ex_med 235629 non-null int64\n",
"mths_since_last_major_derog 66478 non-null float64\n",
"policy_code 235629 non-null int64\n",
"dtypes: float64(16), int64(17), object(23)\n",
"memory usage: 102.5+ MB\n"
]
}
],
"source": [
"df = pd.read_csv(\"LoanStats3c_securev.csv\",skiprows=1)\n",
"df.info()"
]
},
{
"cell_type": "code",
"execution_count": 100,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>id</th>\n",
" <th>member_id</th>\n",
" <th>loan_amnt</th>\n",
" <th>funded_amnt</th>\n",
" <th>funded_amnt_inv</th>\n",
" <th>term</th>\n",
" <th>int_rate</th>\n",
" <th>installment</th>\n",
" <th>grade</th>\n",
" <th>sub_grade</th>\n",
" <th>...</th>\n",
" <th>collection_recovery_fee</th>\n",
" <th>last_pymnt_d</th>\n",
" <th>last_pymnt_amnt</th>\n",
" <th>next_pymnt_d</th>\n",
" <th>last_credit_pull_d</th>\n",
" <th>last_fico_range_high</th>\n",
" <th>last_fico_range_low</th>\n",
" <th>collections_12_mths_ex_med</th>\n",
" <th>mths_since_last_major_derog</th>\n",
" <th>policy_code</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>36805548</td>\n",
" <td>39558264</td>\n",
" <td>10400</td>\n",
" <td>10400</td>\n",
" <td>10400</td>\n",
" <td>36 months</td>\n",
" <td>6.99%</td>\n",
" <td>321.08</td>\n",
" <td>A</td>\n",
" <td>A3</td>\n",
" <td>...</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>321.08</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>679</td>\n",
" <td>675</td>\n",
" <td>0</td>\n",
" <td>59</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>38098114</td>\n",
" <td>40860827</td>\n",
" <td>15000</td>\n",
" <td>15000</td>\n",
" <td>15000</td>\n",
" <td>60 months</td>\n",
" <td>12.39%</td>\n",
" <td>336.64</td>\n",
" <td>C</td>\n",
" <td>C1</td>\n",
" <td>...</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>336.64</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>684</td>\n",
" <td>680</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>37662224</td>\n",
" <td>40425321</td>\n",
" <td>7650</td>\n",
" <td>7650</td>\n",
" <td>7650</td>\n",
" <td>36 months</td>\n",
" <td>13.66%</td>\n",
" <td>260.20</td>\n",
" <td>C</td>\n",
" <td>C3</td>\n",
" <td>...</td>\n",
" <td>0</td>\n",
" <td>May-2015</td>\n",
" <td>260.20</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>689</td>\n",
" <td>685</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>1</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>3 rows × 56 columns</p>\n",
"</div>"
],
"text/plain": [
" id member_id loan_amnt funded_amnt funded_amnt_inv term \\\n",
"0 36805548 39558264 10400 10400 10400 36 months \n",
"1 38098114 40860827 15000 15000 15000 60 months \n",
"2 37662224 40425321 7650 7650 7650 36 months \n",
"\n",
" int_rate installment grade sub_grade ... collection_recovery_fee \\\n",
"0 6.99% 321.08 A A3 ... 0 \n",
"1 12.39% 336.64 C C1 ... 0 \n",
"2 13.66% 260.20 C C3 ... 0 \n",
"\n",
" last_pymnt_d last_pymnt_amnt next_pymnt_d last_credit_pull_d \\\n",
"0 Jun-2015 321.08 Jul-2015 Jun-2015 \n",
"1 Jun-2015 336.64 Jul-2015 Jun-2015 \n",
"2 May-2015 260.20 Jul-2015 Jun-2015 \n",
"\n",
" last_fico_range_high last_fico_range_low collections_12_mths_ex_med \\\n",
"0 679 675 0 \n",
"1 684 680 0 \n",
"2 689 685 0 \n",
"\n",
" mths_since_last_major_derog policy_code \n",
"0 59 1 \n",
"1 NaN 1 \n",
"2 NaN 1 \n",
"\n",
"[3 rows x 56 columns]"
]
},
"execution_count": 100,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.head(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Keeping what we need\n",
"Let's work through our list of attributes slice-by-slice and figure out what we really need. This involves dropping some useless attributes and cleaning-up/modifying others. Let's work through the columns in batches to keep the cognitive burden low:"
]
},
{
"cell_type": "code",
"execution_count": 101,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>id</th>\n",
" <th>member_id</th>\n",
" <th>loan_amnt</th>\n",
" <th>funded_amnt</th>\n",
" <th>funded_amnt_inv</th>\n",
" <th>term</th>\n",
" <th>int_rate</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>36805548</td>\n",
" <td>39558264</td>\n",
" <td>10400</td>\n",
" <td>10400</td>\n",
" <td>10400</td>\n",
" <td>36 months</td>\n",
" <td>6.99%</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>38098114</td>\n",
" <td>40860827</td>\n",
" <td>15000</td>\n",
" <td>15000</td>\n",
" <td>15000</td>\n",
" <td>60 months</td>\n",
" <td>12.39%</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>37662224</td>\n",
" <td>40425321</td>\n",
" <td>7650</td>\n",
" <td>7650</td>\n",
" <td>7650</td>\n",
" <td>36 months</td>\n",
" <td>13.66%</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>37612354</td>\n",
" <td>40375473</td>\n",
" <td>12800</td>\n",
" <td>12800</td>\n",
" <td>12800</td>\n",
" <td>60 months</td>\n",
" <td>17.14%</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>37822187</td>\n",
" <td>40585251</td>\n",
" <td>9600</td>\n",
" <td>9600</td>\n",
" <td>9600</td>\n",
" <td>36 months</td>\n",
" <td>13.66%</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" id member_id loan_amnt funded_amnt funded_amnt_inv term \\\n",
"0 36805548 39558264 10400 10400 10400 36 months \n",
"1 38098114 40860827 15000 15000 15000 60 months \n",
"2 37662224 40425321 7650 7650 7650 36 months \n",
"3 37612354 40375473 12800 12800 12800 60 months \n",
"4 37822187 40585251 9600 9600 9600 36 months \n",
"\n",
" int_rate \n",
"0 6.99% \n",
"1 12.39% \n",
"2 13.66% \n",
"3 17.14% \n",
"4 13.66% "
]
},
"execution_count": 101,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# .ix[row slice, column slice] \n",
"df.ix[:4,:7]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"1. We won't need id or member_id as it has no real predictive power so we can drop them from this table\n",
"2. int_rate was loaded as an `object` data type instead of `float` due to the '%' character. Let's strip that out and convert the column type."
]
},
{
"cell_type": "code",
"execution_count": 102,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.drop(['id','member_id'],1, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 103,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.int_rate = pd.Series(df.int_rate).str.replace('%', '').astype(float)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Loan Amount Requested Verus the Funded Amount \n",
"At a glance I thought it important to discover the relationship between loan amount requested verus the funded amount to see if Lending Club ever issues a lower amount than what is asked and **why**. However after taking a look at the data dictionary I discovered that the data has already been manipulated: \n",
"\n",
">*\"The listed amount of the loan applied for by the borrower. If at some point in time, the credit department reduces the loan amount, then it will be reflected in this value.\"*\n",
"\n",
"Alas, it would have been interesting to see how often Lending Club issues a loan that differs from the requested amount. Doing a quick santiy check below we see there are no instances in the data where the reuqested amount doesn't match the funded."
]
},
{
"cell_type": "code",
"execution_count": 104,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"False 235629\n",
"dtype: int64\n"
]
}
],
"source": [
"print (df.loan_amnt != df.funded_amnt).value_counts() # Instead of old subset syntax let's use the new `query` method.\n",
"\n",
"# df.query('loan_amnt != funded_amnt').ix[:,:5].head(3) # Old way"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's keep moving through the columns"
]
},
{
"cell_type": "code",
"execution_count": 105,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>emp_title</th>\n",
" <th>emp_length</th>\n",
" <th>home_ownership</th>\n",
" <th>annual_inc</th>\n",
" <th>verification_status</th>\n",
" <th>issue_d</th>\n",
" <th>loan_status</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Truck Driver Delivery Personel</td>\n",
" <td>8 years</td>\n",
" <td>MORTGAGE</td>\n",
" <td>58000</td>\n",
" <td>not verified</td>\n",
" <td>Dec-2014</td>\n",
" <td>Current</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>MANAGEMENT</td>\n",
" <td>10+ years</td>\n",
" <td>RENT</td>\n",
" <td>78000</td>\n",
" <td>VERIFIED - income source</td>\n",
" <td>Dec-2014</td>\n",
" <td>Current</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Technical Specialist</td>\n",
" <td>&lt; 1 year</td>\n",
" <td>RENT</td>\n",
" <td>50000</td>\n",
" <td>VERIFIED - income source</td>\n",
" <td>Dec-2014</td>\n",
" <td>Late (31-120 days)</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Senior Sales Professional</td>\n",
" <td>10+ years</td>\n",
" <td>MORTGAGE</td>\n",
" <td>125000</td>\n",
" <td>VERIFIED - income</td>\n",
" <td>Dec-2014</td>\n",
" <td>Current</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>Admin Specialist</td>\n",
" <td>10+ years</td>\n",
" <td>RENT</td>\n",
" <td>69000</td>\n",
" <td>VERIFIED - income source</td>\n",
" <td>Dec-2014</td>\n",
" <td>Fully Paid</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>Programming Analysis Supervisor</td>\n",
" <td>6 years</td>\n",
" <td>RENT</td>\n",
" <td>63800</td>\n",
" <td>VERIFIED - income source</td>\n",
" <td>Dec-2014</td>\n",
" <td>Current</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" emp_title emp_length home_ownership annual_inc \\\n",
"0 Truck Driver Delivery Personel 8 years MORTGAGE 58000 \n",
"1 MANAGEMENT 10+ years RENT 78000 \n",
"2 Technical Specialist < 1 year RENT 50000 \n",
"3 Senior Sales Professional 10+ years MORTGAGE 125000 \n",
"4 Admin Specialist 10+ years RENT 69000 \n",
"5 Programming Analysis Supervisor 6 years RENT 63800 \n",
"\n",
" verification_status issue_d loan_status \n",
"0 not verified Dec-2014 Current \n",
"1 VERIFIED - income source Dec-2014 Current \n",
"2 VERIFIED - income source Dec-2014 Late (31-120 days) \n",
"3 VERIFIED - income Dec-2014 Current \n",
"4 VERIFIED - income source Dec-2014 Fully Paid \n",
"5 VERIFIED - income source Dec-2014 Current "
]
},
"execution_count": 105,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,8:15]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Employment Title\n",
"`emp_title` might be a free text field on the application form or a list of currated employment titles. Let's examine how many unique values exist:"
]
},
{
"cell_type": "code",
"execution_count": 106,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Teacher 4569\n",
"Manager 3772\n",
"Registered Nurse 1960\n",
"RN 1816\n",
"Supervisor 1663\n",
"Name: emp_title, dtype: int64\n",
"Teacher/Media Specialist 1\n",
"teacher/ After School Director 1\n",
"bpd/gms coordinator 1\n",
"heavy equip operator 1\n",
"Corp. Operations Mangager 1\n",
"Name: emp_title, dtype: int64\n"
]
},
{
"data": {
"text/plain": [
"(75353,)"
]
},
"execution_count": 106,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print df.emp_title.value_counts().head()\n",
"print df.emp_title.value_counts().tail()\n",
"df.emp_title.unique().shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"75,353 unique entries among 235,630 observations seems like a bit of a stretch. Taking a look at the head vs tail of the doc shows some suspiciously specific titles such as `bpd/gms coordinator`. I feel comfortable assessing that this data won't be meaningful and any relationship we might observe might be due to confounding relationships. A more advanced implementation might look to group all these job descriptions into categories and/or examine if Lending Club's model looks at (income + job) versus just income, but that's out of the scope of this post.\n",
"\n",
"Here we define a confounding variable as a variable that obscures the effects of another variable. If one 1st grade teacher used a phonics textbook in her class and a different teacher used a whole language textbook in his class, and students in the two classes were\n",
"given achievement tests to see how well they read, the independent variables\n",
"(teacher effectiveness and textbooks) would be confounded. There is no way to\n",
"determine if differences in reading between the two classes were caused by\n",
"either or both of the independent variables.\n",
"\n",
"Applying this example to our dataset, Registered Nurses, or RNs, who have higher education requirements and receive above average pay, might be granted A grade loans on average. Is this due to them working as RNs or having a 4-year degree or their salary? Would we see the same effect for Physician Assistants who go to school for longer and receive even more pay? What about Certified Nursing Assistants (CNAs) who are on the oppposite spectrum?"
]
},
{
"cell_type": "code",
"execution_count": 107,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"df.drop(['emp_title'],1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Employment Length\n",
"Leaving this variable in might contradict our decision to drop the employment tile as it also conveyed a sort of socio-economic seniority. A Computer Scientist 5 years into their career would generally have a larger salary than a Kindergarden teacher 10 years into their career. Arguably it might be powerful to combine a grouped, matched, and reduced set of employment titles with their length to create a \"purchasing power\" metric. Since employment length is an easy scalar, let's leave it in for now. \n",
"\n",
"We could leave 'emp_length' as categorical data, but it shouldn't be treated as such or as ordinal data since the intervals are easy to determine. We can convert it into numerical data with a simple filter:"
]
},
{
"cell_type": "code",
"execution_count": 108,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"data": {
"text/plain": [
"10+ years 79505\n",
"2 years 20487\n",
"3 years 18267\n",
"< 1 year 17982\n",
"1 year 14593\n",
"4 years 13528\n",
"7 years 13099\n",
"5 years 13051\n",
"n/a 12019\n",
"8 years 11853\n",
"6 years 11821\n",
"9 years 9424\n",
"Name: emp_length, dtype: int64"
]
},
"execution_count": 108,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.emp_length.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 109,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.replace('n/a', np.nan,inplace=True)\n",
"df.emp_length.fillna(value=0,inplace=True)\n",
"df['emp_length'].replace(to_replace='[^0-9]+', value='', inplace=True, regex=True)\n",
"df['emp_length'] = df['emp_length'].astype(int)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Verification Status\n",
"The values of `verification_status` aren't immediately clear. As `VERIFIED - income` and `VERIFIED - income source` could potentially be the same criteria. The provided data dictionary file describes the field as such: _\"Indicates if income was verified by LC, not verified, or if the income source was verified\"_ This description does not help as you'd intuitively think that you would have to verify the source in order to verify income. For now we'll trust that these are indeed different methods or levels (perhaps even ordinal) of verification."
]
},
{
"cell_type": "code",
"execution_count": 110,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"VERIFIED - income source 97740\n",
"not verified 70661\n",
"VERIFIED - income 67228\n",
"Name: verification_status, dtype: int64"
]
},
"execution_count": 110,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.verification_status.value_counts()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Loan Statuses\n",
"Moving on to the last column in the slice of the df we are examining: loan status is mutable value that represents the current state of the loan. If anything we might want to examine if all the independent variables and/or interst rate to determine the probability of the loan status at some time. In this post however we are focused on predicting the interest rate granted to an applicant at loan creation. Thus we do not care about this variable and will drop it after examining it out of curiosity.\n",
"\n",
"When examining the distribution among the loan statuses, October has the highest amount of loans right ahead of the holiday season. To clean up the space let's replace the column value with a month only and convert the column back to a string or object type. "
]
},
{
"cell_type": "code",
"execution_count": 111,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Current 194161\n",
"Fully Paid 29346\n",
"Charged Off 5488\n",
"Late (31-120 days) 3738\n",
"In Grace Period 1996\n",
"Late (16-30 days) 824\n",
"Default 76\n",
"Name: loan_status, dtype: int64\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAaUAAAHeCAYAAAA7NxSsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKZMdQAAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x11145b350>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"print df.loan_status.value_counts()\n",
"\n",
"issue_d_todate = pd.to_datetime(df.issue_d)# (df['issue_d'].apply(lambda x: x.strftime('%Y-%m-%d')))\n",
"df.issue_d = pd.Series(df.issue_d).str.replace('-2014', '')\n",
"# We need sort_index() or else we won't get a sequential timedate order.\n",
"issue_d_todate.value_counts().sort_index().plot(kind='bar')\n",
"\n",
"df.drop(['loan_status'],1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Arbitrary Applicant Input\n",
"Continuing on to examine more columns:"
]
},
{
"cell_type": "code",
"execution_count": 112,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>issue_d</th>\n",
" <th>pymnt_plan</th>\n",
" <th>url</th>\n",
" <th>desc</th>\n",
" <th>purpose</th>\n",
" <th>title</th>\n",
" <th>zip_code</th>\n",
" <th>addr_state</th>\n",
" <th>dti</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>credit_card</td>\n",
" <td>Credit card refinancing</td>\n",
" <td>937xx</td>\n",
" <td>CA</td>\n",
" <td>14.92</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>debt_consolidation</td>\n",
" <td>Debt consolidation</td>\n",
" <td>235xx</td>\n",
" <td>VA</td>\n",
" <td>12.03</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>debt_consolidation</td>\n",
" <td>Debt consolidation</td>\n",
" <td>850xx</td>\n",
" <td>AZ</td>\n",
" <td>34.81</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>car</td>\n",
" <td>Car financing</td>\n",
" <td>953xx</td>\n",
" <td>CA</td>\n",
" <td>8.31</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>debt_consolidation</td>\n",
" <td>Debt consolidation</td>\n",
" <td>077xx</td>\n",
" <td>NJ</td>\n",
" <td>25.81</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>Dec</td>\n",
" <td>n</td>\n",
" <td>https://www.lendingclub.com/browse/loanDetail....</td>\n",
" <td>NaN</td>\n",
" <td>credit_card</td>\n",
" <td>Credit card refinancing</td>\n",
" <td>658xx</td>\n",
" <td>MO</td>\n",
" <td>18.49</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" issue_d pymnt_plan url desc \\\n",
"0 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"1 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"2 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"3 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"4 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"5 Dec n https://www.lendingclub.com/browse/loanDetail.... NaN \n",
"\n",
" purpose title zip_code addr_state dti \n",
"0 credit_card Credit card refinancing 937xx CA 14.92 \n",
"1 debt_consolidation Debt consolidation 235xx VA 12.03 \n",
"2 debt_consolidation Debt consolidation 850xx AZ 34.81 \n",
"3 car Car financing 953xx CA 8.31 \n",
"4 debt_consolidation Debt consolidation 077xx NJ 25.81 \n",
"5 credit_card Credit card refinancing 658xx MO 18.49 "
]
},
"execution_count": 112,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,12:21]"
]
},
{
"cell_type": "code",
"execution_count": 113,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"debt_consolidation 143006\n",
"credit_card 55522\n",
"home_improvement 13045\n",
"other 10371\n",
"major_purchase 3858\n",
"medical 2331\n",
"small_business 2277\n",
"car 1832\n",
"moving 1328\n",
"vacation 1178\n",
"house 750\n",
"renewable_energy 123\n",
"wedding 8\n",
"Name: purpose, dtype: int64\n",
"\n",
"consolodate 1\n",
"Chase debt 1\n",
"pay off credit cards and loans 1\n",
"Business Consolidation Loan 1\n",
"Bills Bills Bills 1\n",
"Name: title, dtype: int64\n"
]
}
],
"source": [
"print df.purpose.value_counts()\n",
"print ''\n",
"print df.title.value_counts().tail()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We'll be dropping `pymnt_plan` as it has the same current state info that `loan_status` had above. It indicates that the loan is in jeopardy and that the borrower has been placed on a payment plan.\n",
"\n",
"Examining the unique counts of purpose show that they are lending club selected options;howver, `title` and `desc` are arbitrary free-text from the applicant."
]
},
{
"cell_type": "code",
"execution_count": 114,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.drop(['pymnt_plan','url','desc','title' ],1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Moving on to examine the next slice of columns:"
]
},
{
"cell_type": "code",
"execution_count": 115,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>delinq_2yrs</th>\n",
" <th>earliest_cr_line</th>\n",
" <th>fico_range_low</th>\n",
" <th>fico_range_high</th>\n",
" <th>inq_last_6mths</th>\n",
" <th>mths_since_last_delinq</th>\n",
" <th>mths_since_last_record</th>\n",
" <th>open_acc</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>Sep-1989</td>\n",
" <td>710</td>\n",
" <td>714</td>\n",
" <td>2</td>\n",
" <td>42</td>\n",
" <td>NaN</td>\n",
" <td>17</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0</td>\n",
" <td>Aug-1994</td>\n",
" <td>750</td>\n",
" <td>754</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>6</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>0</td>\n",
" <td>Aug-2002</td>\n",
" <td>685</td>\n",
" <td>689</td>\n",
" <td>1</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>11</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>1</td>\n",
" <td>Oct-2000</td>\n",
" <td>665</td>\n",
" <td>669</td>\n",
" <td>0</td>\n",
" <td>17</td>\n",
" <td>NaN</td>\n",
" <td>8</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0</td>\n",
" <td>Nov-1992</td>\n",
" <td>680</td>\n",
" <td>684</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>12</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>0</td>\n",
" <td>Aug-2003</td>\n",
" <td>685</td>\n",
" <td>689</td>\n",
" <td>0</td>\n",
" <td>60</td>\n",
" <td>NaN</td>\n",
" <td>10</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" delinq_2yrs earliest_cr_line fico_range_low fico_range_high \\\n",
"0 0 Sep-1989 710 714 \n",
"1 0 Aug-1994 750 754 \n",
"2 0 Aug-2002 685 689 \n",
"3 1 Oct-2000 665 669 \n",
"4 0 Nov-1992 680 684 \n",
"5 0 Aug-2003 685 689 \n",
"\n",
" inq_last_6mths mths_since_last_delinq mths_since_last_record open_acc \n",
"0 2 42 NaN 17 \n",
"1 0 NaN NaN 6 \n",
"2 1 NaN NaN 11 \n",
"3 0 17 NaN 8 \n",
"4 0 NaN NaN 12 \n",
"5 0 60 NaN 10 "
]
},
"execution_count": 115,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,17:25]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Delinquency \n",
"Lets examine the distribution of delinquency across all applicants:"
]
},
{
"cell_type": "code",
"execution_count": 116,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"0 186781\n",
"1 31859\n",
"2 9831\n",
"3 3654\n",
"4 1606\n",
"5 849\n",
"6 426\n",
"7 254\n",
"8 130\n",
"9 84\n",
"10 47\n",
"12 30\n",
"11 28\n",
"13 21\n",
"14 7\n",
"15 5\n",
"16 5\n",
"18 3\n",
"19 3\n",
"21 2\n",
"17 2\n",
"22 2\n",
"Name: delinq_2yrs, dtype: int64"
]
},
"execution_count": 116,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.delinq_2yrs.value_counts()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Earliest Credit Line\n",
"`earliest_cr_line` is currently an object dtype however we don't want to convert it to a date, rather a scalar to describe the length of time since the first line of credit. We are asserting that if all variables are held equal (number of lines open, income, delinquencies, verfied income); the longer you've had lines of credit the better. \n",
"\n",
"Conversely an acturary might asses that we can end up with a \"u-shaped\" distribution in that if you are the extreme end of years, you are a higher risk as you have a higher chance of mortality thus a lower probability of repayment. Additionally pre-retirement debitors are more likely to list job or health reasons as the primary cause of their bankruptcy. \n",
"\n",
"In a more advanced implementation we'd want to account for confounding variables in that certain applicant groups, the older you are (time since first credit line) the larger your earning potential/power and thus the better loan you might secure, however this increase in salary might be negligible if the amount of discretionary money spent or need for a loan scales in some proportion to the salary size."
]
},
{
"cell_type": "code",
"execution_count": 117,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"0 26\n",
"1 21\n",
"2 13\n",
"3 15\n",
"4 23\n",
"5 12\n",
"6 16\n",
"7 17\n",
"8 14\n",
"9 12\n",
"10 14\n",
"11 6\n",
"12 28\n",
"13 12\n",
"14 20\n",
"15 21\n",
"16 19\n",
"17 26\n",
"18 14\n",
"19 17\n",
"20 9\n",
"21 13\n",
"22 14\n",
"23 8\n",
"24 5\n",
"25 10\n",
"26 13\n",
"27 13\n",
"28 8\n",
"29 21\n",
" ..\n",
"235599 15\n",
"235600 14\n",
"235601 30\n",
"235602 22\n",
"235603 30\n",
"235604 18\n",
"235605 18\n",
"235606 22\n",
"235607 12\n",
"235608 17\n",
"235609 22\n",
"235610 21\n",
"235611 13\n",
"235612 13\n",
"235613 29\n",
"235614 15\n",
"235615 21\n",
"235616 25\n",
"235617 14\n",
"235618 32\n",
"235619 21\n",
"235620 22\n",
"235621 17\n",
"235622 13\n",
"235623 16\n",
"235624 12\n",
"235625 18\n",
"235626 12\n",
"235627 13\n",
"235628 15\n",
"Name: earliest_cr_line, dtype: int64"
]
},
"execution_count": 117,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from datetime import datetime\n",
"\n",
"df.earliest_cr_line = pd.to_datetime(df.earliest_cr_line)\n",
"\n",
"dttoday = datetime.now().strftime('%Y-%m-%d')\n",
"# There is a better way to do this :) \n",
"df.earliest_cr_line = df.earliest_cr_line.apply(lambda x: (\n",
" np.timedelta64((x - pd.Timestamp(dttoday)),'D').astype(int))/-365)\n",
"\n",
"df.earliest_cr_line"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## FICO Ranges\n",
"The FICO `fico_range_low` & `fico_range_high` scores on their own aren't as useful as a range and average. So let's create that. `initial_list_status` is unclear if it describes a post or pre determined interest period. "
]
},
{
"cell_type": "code",
"execution_count": 118,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df['fico_range'] = df.fico_range_low.astype('str') + '-' + df.fico_range_high.astype('str')\n",
"df['meanfico'] = (df.fico_range_low + df.fico_range_high)/2\n",
"df.drop(['fico_range_low','fico_range_high','initial_list_status'],1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Moving on to the next set of columns"
]
},
{
"cell_type": "code",
"execution_count": 119,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>pub_rec</th>\n",
" <th>revol_bal</th>\n",
" <th>revol_util</th>\n",
" <th>total_acc</th>\n",
" <th>out_prncp</th>\n",
" <th>out_prncp_inv</th>\n",
" <th>total_pymnt</th>\n",
" <th>total_pymnt_inv</th>\n",
" <th>total_rec_prncp</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>6133</td>\n",
" <td>31.6%</td>\n",
" <td>36</td>\n",
" <td>9082.24</td>\n",
" <td>9082.24</td>\n",
" <td>1595.30</td>\n",
" <td>1595.30</td>\n",
" <td>1317.76</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0</td>\n",
" <td>138008</td>\n",
" <td>29%</td>\n",
" <td>17</td>\n",
" <td>14072.21</td>\n",
" <td>14072.21</td>\n",
" <td>1657.39</td>\n",
" <td>1657.39</td>\n",
" <td>927.79</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>0</td>\n",
" <td>16822</td>\n",
" <td>91.9%</td>\n",
" <td>20</td>\n",
" <td>6945.62</td>\n",
" <td>6945.62</td>\n",
" <td>1026.29</td>\n",
" <td>1026.29</td>\n",
" <td>704.38</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>0</td>\n",
" <td>5753</td>\n",
" <td>100.9%</td>\n",
" <td>13</td>\n",
" <td>12098.99</td>\n",
" <td>12098.99</td>\n",
" <td>1564.93</td>\n",
" <td>1564.93</td>\n",
" <td>701.01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0</td>\n",
" <td>16388</td>\n",
" <td>59.4%</td>\n",
" <td>44</td>\n",
" <td>0.00</td>\n",
" <td>0.00</td>\n",
" <td>9973.43</td>\n",
" <td>9973.43</td>\n",
" <td>9600.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>0</td>\n",
" <td>16374</td>\n",
" <td>76.2%</td>\n",
" <td>35</td>\n",
" <td>20203.61</td>\n",
" <td>20203.61</td>\n",
" <td>2535.41</td>\n",
" <td>2535.41</td>\n",
" <td>1221.39</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" pub_rec revol_bal revol_util total_acc out_prncp out_prncp_inv \\\n",
"0 0 6133 31.6% 36 9082.24 9082.24 \n",
"1 0 138008 29% 17 14072.21 14072.21 \n",
"2 0 16822 91.9% 20 6945.62 6945.62 \n",
"3 0 5753 100.9% 13 12098.99 12098.99 \n",
"4 0 16388 59.4% 44 0.00 0.00 \n",
"5 0 16374 76.2% 35 20203.61 20203.61 \n",
"\n",
" total_pymnt total_pymnt_inv total_rec_prncp \n",
"0 1595.30 1595.30 1317.76 \n",
"1 1657.39 1657.39 927.79 \n",
"2 1026.29 1026.29 704.38 \n",
"3 1564.93 1564.93 701.01 \n",
"4 9973.43 9973.43 9600.00 \n",
"5 2535.41 2535.41 1221.39 "
]
},
"execution_count": 119,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,23:32]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Revolving Utility\n",
"Let's do the same thing we did for the interest rate column"
]
},
{
"cell_type": "code",
"execution_count": 120,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"df.revol_util = pd.Series(df.revol_util).str.replace('%', '').astype(float)"
]
},
{
"cell_type": "code",
"execution_count": 121,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>pub_rec</th>\n",
" <th>revol_bal</th>\n",
" <th>revol_util</th>\n",
" <th>total_acc</th>\n",
" <th>out_prncp</th>\n",
" <th>out_prncp_inv</th>\n",
" <th>total_pymnt</th>\n",
" <th>total_pymnt_inv</th>\n",
" <th>total_rec_prncp</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>6133</td>\n",
" <td>31.6</td>\n",
" <td>36</td>\n",
" <td>9082.24</td>\n",
" <td>9082.24</td>\n",
" <td>1595.30</td>\n",
" <td>1595.30</td>\n",
" <td>1317.76</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0</td>\n",
" <td>138008</td>\n",
" <td>29.0</td>\n",
" <td>17</td>\n",
" <td>14072.21</td>\n",
" <td>14072.21</td>\n",
" <td>1657.39</td>\n",
" <td>1657.39</td>\n",
" <td>927.79</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>0</td>\n",
" <td>16822</td>\n",
" <td>91.9</td>\n",
" <td>20</td>\n",
" <td>6945.62</td>\n",
" <td>6945.62</td>\n",
" <td>1026.29</td>\n",
" <td>1026.29</td>\n",
" <td>704.38</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>0</td>\n",
" <td>5753</td>\n",
" <td>100.9</td>\n",
" <td>13</td>\n",
" <td>12098.99</td>\n",
" <td>12098.99</td>\n",
" <td>1564.93</td>\n",
" <td>1564.93</td>\n",
" <td>701.01</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0</td>\n",
" <td>16388</td>\n",
" <td>59.4</td>\n",
" <td>44</td>\n",
" <td>0.00</td>\n",
" <td>0.00</td>\n",
" <td>9973.43</td>\n",
" <td>9973.43</td>\n",
" <td>9600.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>0</td>\n",
" <td>16374</td>\n",
" <td>76.2</td>\n",
" <td>35</td>\n",
" <td>20203.61</td>\n",
" <td>20203.61</td>\n",
" <td>2535.41</td>\n",
" <td>2535.41</td>\n",
" <td>1221.39</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" pub_rec revol_bal revol_util total_acc out_prncp out_prncp_inv \\\n",
"0 0 6133 31.6 36 9082.24 9082.24 \n",
"1 0 138008 29.0 17 14072.21 14072.21 \n",
"2 0 16822 91.9 20 6945.62 6945.62 \n",
"3 0 5753 100.9 13 12098.99 12098.99 \n",
"4 0 16388 59.4 44 0.00 0.00 \n",
"5 0 16374 76.2 35 20203.61 20203.61 \n",
"\n",
" total_pymnt total_pymnt_inv total_rec_prncp \n",
"0 1595.30 1595.30 1317.76 \n",
"1 1657.39 1657.39 927.79 \n",
"2 1026.29 1026.29 704.38 \n",
"3 1564.93 1564.93 701.01 \n",
"4 9973.43 9973.43 9600.00 \n",
"5 2535.41 2535.41 1221.39 "
]
},
"execution_count": 121,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,23:32]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## More Post Loan Attributes\n",
"Similar to some attributes we encountered above, we have more columns to drop that detail attributes about the loan after it was granted, which we don't care about."
]
},
{
"cell_type": "code",
"execution_count": 122,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.drop(['out_prncp','out_prncp_inv','total_pymnt',\n",
" 'total_pymnt_inv','total_rec_prncp', 'grade', 'sub_grade'] ,1, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 123,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>revol_util</th>\n",
" <th>total_acc</th>\n",
" <th>total_rec_int</th>\n",
" <th>total_rec_late_fee</th>\n",
" <th>recoveries</th>\n",
" <th>collection_recovery_fee</th>\n",
" <th>last_pymnt_d</th>\n",
" <th>last_pymnt_amnt</th>\n",
" <th>next_pymnt_d</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>31.6</td>\n",
" <td>36</td>\n",
" <td>277.54</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>321.08</td>\n",
" <td>Jul-2015</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>29.0</td>\n",
" <td>17</td>\n",
" <td>729.60</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>336.64</td>\n",
" <td>Jul-2015</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>91.9</td>\n",
" <td>20</td>\n",
" <td>321.91</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>May-2015</td>\n",
" <td>260.20</td>\n",
" <td>Jul-2015</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>100.9</td>\n",
" <td>13</td>\n",
" <td>863.92</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>319.08</td>\n",
" <td>Jul-2015</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>59.4</td>\n",
" <td>44</td>\n",
" <td>373.43</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>Apr-2015</td>\n",
" <td>9338.58</td>\n",
" <td>NaN</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>76.2</td>\n",
" <td>35</td>\n",
" <td>1314.02</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>Jun-2015</td>\n",
" <td>516.36</td>\n",
" <td>Jul-2015</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" revol_util total_acc total_rec_int total_rec_late_fee recoveries \\\n",
"0 31.6 36 277.54 0 0 \n",
"1 29.0 17 729.60 0 0 \n",
"2 91.9 20 321.91 0 0 \n",
"3 100.9 13 863.92 0 0 \n",
"4 59.4 44 373.43 0 0 \n",
"5 76.2 35 1314.02 0 0 \n",
"\n",
" collection_recovery_fee last_pymnt_d last_pymnt_amnt next_pymnt_d \n",
"0 0 Jun-2015 321.08 Jul-2015 \n",
"1 0 Jun-2015 336.64 Jul-2015 \n",
"2 0 May-2015 260.20 Jul-2015 \n",
"3 0 Jun-2015 319.08 Jul-2015 \n",
"4 0 Apr-2015 9338.58 NaN \n",
"5 0 Jun-2015 516.36 Jul-2015 "
]
},
"execution_count": 123,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,23:32]"
]
},
{
"cell_type": "code",
"execution_count": 124,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"df.drop(['total_rec_int','total_rec_late_fee',\n",
" 'recoveries','collection_recovery_fee',\n",
" 'collection_recovery_fee' ],1, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 125,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>revol_util</th>\n",
" <th>total_acc</th>\n",
" <th>last_pymnt_d</th>\n",
" <th>last_pymnt_amnt</th>\n",
" <th>next_pymnt_d</th>\n",
" <th>last_credit_pull_d</th>\n",
" <th>last_fico_range_high</th>\n",
" <th>last_fico_range_low</th>\n",
" <th>collections_12_mths_ex_med</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>31.6</td>\n",
" <td>36</td>\n",
" <td>Jun-2015</td>\n",
" <td>321.08</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>679</td>\n",
" <td>675</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>29.0</td>\n",
" <td>17</td>\n",
" <td>Jun-2015</td>\n",
" <td>336.64</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>684</td>\n",
" <td>680</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>91.9</td>\n",
" <td>20</td>\n",
" <td>May-2015</td>\n",
" <td>260.20</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>689</td>\n",
" <td>685</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>100.9</td>\n",
" <td>13</td>\n",
" <td>Jun-2015</td>\n",
" <td>319.08</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>684</td>\n",
" <td>680</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>59.4</td>\n",
" <td>44</td>\n",
" <td>Apr-2015</td>\n",
" <td>9338.58</td>\n",
" <td>NaN</td>\n",
" <td>Apr-2015</td>\n",
" <td>679</td>\n",
" <td>675</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>76.2</td>\n",
" <td>35</td>\n",
" <td>Jun-2015</td>\n",
" <td>516.36</td>\n",
" <td>Jul-2015</td>\n",
" <td>Jun-2015</td>\n",
" <td>689</td>\n",
" <td>685</td>\n",
" <td>0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" revol_util total_acc last_pymnt_d last_pymnt_amnt next_pymnt_d \\\n",
"0 31.6 36 Jun-2015 321.08 Jul-2015 \n",
"1 29.0 17 Jun-2015 336.64 Jul-2015 \n",
"2 91.9 20 May-2015 260.20 Jul-2015 \n",
"3 100.9 13 Jun-2015 319.08 Jul-2015 \n",
"4 59.4 44 Apr-2015 9338.58 NaN \n",
"5 76.2 35 Jun-2015 516.36 Jul-2015 \n",
"\n",
" last_credit_pull_d last_fico_range_high last_fico_range_low \\\n",
"0 Jun-2015 679 675 \n",
"1 Jun-2015 684 680 \n",
"2 Jun-2015 689 685 \n",
"3 Jun-2015 684 680 \n",
"4 Apr-2015 679 675 \n",
"5 Jun-2015 689 685 \n",
"\n",
" collections_12_mths_ex_med \n",
"0 0 \n",
"1 0 \n",
"2 0 \n",
"3 0 \n",
"4 0 \n",
"5 0 "
]
},
"execution_count": 125,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,23:32]"
]
},
{
"cell_type": "code",
"execution_count": 126,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.drop(['last_pymnt_d','last_pymnt_amnt',\n",
" 'next_pymnt_d','last_credit_pull_d'],1, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 127,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>revol_util</th>\n",
" <th>total_acc</th>\n",
" <th>last_fico_range_high</th>\n",
" <th>last_fico_range_low</th>\n",
" <th>collections_12_mths_ex_med</th>\n",
" <th>mths_since_last_major_derog</th>\n",
" <th>policy_code</th>\n",
" <th>fico_range</th>\n",
" <th>meanfico</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>31.6</td>\n",
" <td>36</td>\n",
" <td>679</td>\n",
" <td>675</td>\n",
" <td>0</td>\n",
" <td>59</td>\n",
" <td>1</td>\n",
" <td>710-714</td>\n",
" <td>712</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>29.0</td>\n",
" <td>17</td>\n",
" <td>684</td>\n",
" <td>680</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>1</td>\n",
" <td>750-754</td>\n",
" <td>752</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>91.9</td>\n",
" <td>20</td>\n",
" <td>689</td>\n",
" <td>685</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>1</td>\n",
" <td>685-689</td>\n",
" <td>687</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>100.9</td>\n",
" <td>13</td>\n",
" <td>684</td>\n",
" <td>680</td>\n",
" <td>0</td>\n",
" <td>36</td>\n",
" <td>1</td>\n",
" <td>665-669</td>\n",
" <td>667</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>59.4</td>\n",
" <td>44</td>\n",
" <td>679</td>\n",
" <td>675</td>\n",
" <td>0</td>\n",
" <td>NaN</td>\n",
" <td>1</td>\n",
" <td>680-684</td>\n",
" <td>682</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>76.2</td>\n",
" <td>35</td>\n",
" <td>689</td>\n",
" <td>685</td>\n",
" <td>0</td>\n",
" <td>74</td>\n",
" <td>1</td>\n",
" <td>685-689</td>\n",
" <td>687</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" revol_util total_acc last_fico_range_high last_fico_range_low \\\n",
"0 31.6 36 679 675 \n",
"1 29.0 17 684 680 \n",
"2 91.9 20 689 685 \n",
"3 100.9 13 684 680 \n",
"4 59.4 44 679 675 \n",
"5 76.2 35 689 685 \n",
"\n",
" collections_12_mths_ex_med mths_since_last_major_derog policy_code \\\n",
"0 0 59 1 \n",
"1 0 NaN 1 \n",
"2 0 NaN 1 \n",
"3 0 36 1 \n",
"4 0 NaN 1 \n",
"5 0 74 1 \n",
"\n",
" fico_range meanfico \n",
"0 710-714 712 \n",
"1 750-754 752 \n",
"2 685-689 687 \n",
"3 665-669 667 \n",
"4 680-684 682 \n",
"5 685-689 687 "
]
},
"execution_count": 127,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.ix[:5,23:32]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Last FICO Ranges\n",
"Our previous FICO attributes were the current ranges."
]
},
{
"cell_type": "code",
"execution_count": 128,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df['last_fico_range'] = df.last_fico_range_low.astype('str') + '-' + df.last_fico_range_high.astype('str')\n",
"df['last_meanfico'] = (df.last_fico_range_low + df.last_fico_range_high)/2\n",
"df.drop(['last_fico_range_high','last_fico_range_low','policy_code'],1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Wrapping Up the Data Munging\n",
"\n",
"Okay so now we got out data down to 61mb from 120mb. Imagine what we could do if we used more efficicent data types! Let's take one last look at it to then move on to our modeling phase."
]
},
{
"cell_type": "code",
"execution_count": 129,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Index([u'loan_amnt', u'funded_amnt', u'funded_amnt_inv', u'term', u'int_rate',\n",
" u'installment', u'emp_length', u'home_ownership', u'annual_inc',\n",
" u'verification_status', u'issue_d', u'purpose', u'zip_code',\n",
" u'addr_state', u'dti', u'delinq_2yrs', u'earliest_cr_line',\n",
" u'inq_last_6mths', u'mths_since_last_delinq', u'mths_since_last_record',\n",
" u'open_acc', u'pub_rec', u'revol_bal', u'revol_util', u'total_acc',\n",
" u'collections_12_mths_ex_med', u'mths_since_last_major_derog',\n",
" u'fico_range', u'meanfico', u'last_fico_range', u'last_meanfico'],\n",
" dtype='object')\n",
"[[10400 10400 10400 ' 36 months' 6.99 321.08 8 'MORTGAGE' 58000.0\n",
" 'not verified' 'Dec' 'credit_card' '937xx' 'CA' 14.92 0 26 2 42.0 nan 17\n",
" 0 6133 31.6 36 0 59.0 '710-714' 712.0 '675-679' 677.0]]\n",
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 235629 entries, 0 to 235628\n",
"Data columns (total 31 columns):\n",
"loan_amnt 235629 non-null int64\n",
"funded_amnt 235629 non-null int64\n",
"funded_amnt_inv 235629 non-null int64\n",
"term 235629 non-null object\n",
"int_rate 235629 non-null float64\n",
"installment 235629 non-null float64\n",
"emp_length 235629 non-null int64\n",
"home_ownership 235629 non-null object\n",
"annual_inc 235629 non-null float64\n",
"verification_status 235629 non-null object\n",
"issue_d 235629 non-null object\n",
"purpose 235629 non-null object\n",
"zip_code 235629 non-null object\n",
"addr_state 235629 non-null object\n",
"dti 235629 non-null float64\n",
"delinq_2yrs 235629 non-null int64\n",
"earliest_cr_line 235629 non-null int64\n",
"inq_last_6mths 235629 non-null int64\n",
"mths_since_last_delinq 119748 non-null float64\n",
"mths_since_last_record 41524 non-null float64\n",
"open_acc 235629 non-null int64\n",
"pub_rec 235629 non-null int64\n",
"revol_bal 235629 non-null int64\n",
"revol_util 235504 non-null float64\n",
"total_acc 235629 non-null int64\n",
"collections_12_mths_ex_med 235629 non-null int64\n",
"mths_since_last_major_derog 66478 non-null float64\n",
"fico_range 235629 non-null object\n",
"meanfico 235629 non-null float64\n",
"last_fico_range 235629 non-null object\n",
"last_meanfico 235629 non-null float64\n",
"dtypes: float64(10), int64(12), object(9)\n",
"memory usage: 57.5+ MB\n"
]
}
],
"source": [
"print df.columns\n",
"print df.head(1).values\n",
"df.info()"
]
},
{
"cell_type": "code",
"execution_count": 130,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.fillna(0.0,inplace=True)\n",
"df.fillna(0,inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# 2. Our model\n",
"Staying true to the original intent of this blog, reproducing my popular old R post, I will use a gradient boosting regression model (more on this later), even if it's easier to use random forest and avoid dummy encoding of my categorical data. If you walk through the broken out steps above you'll find that we have:\n",
"\n",
"1. Addressed missing data\n",
"2. Converted strings to numerical representations where possible\n",
"3. Dropped superfluous attributes\n",
"\n",
"**Note:** We have to pay special attention to how we replace missing values since we can introduce bias if the data is not missing at random.\n",
"\n",
"The usual next steps are to:\n",
"1. Eliminate zero (or near-zero) variance predictors\n",
"2. Highly and correlated predictors\n",
"\n",
"Due to all the categorical variables we might want to use Kendall's correlation matrix. If our data was larger we might want to reduce the size of it via more advanced feature selection methods such as clustering (choosing a single representative of each cluster), PCA, ICA, and so on. Since the purpose of this experiment is explanatory rather than predictive, we wouldn't want to use PCA and try to explain the covariates in terms of principal components.\n",
"\n",
"## Highly Correlated Data\n",
"Let's examine our dataframes correlation matrix and drop highly correlated/redundant data to address multicollinearity."
]
},
{
"cell_type": "code",
"execution_count": 131,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"funded_amnt loan_amnt 1.000000\n",
"funded_amnt_inv loan_amnt 0.999997\n",
" funded_amnt 0.999997\n",
"installment loan_amnt 0.947978\n",
" funded_amnt 0.947978\n",
" funded_amnt_inv 0.947943\n",
"pub_rec mths_since_last_record 0.665077\n",
"total_acc open_acc 0.684919\n",
"dtype: float64"
]
},
"execution_count": 131,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"cor = df.corr()\n",
"cor.loc[:,:] = np.tril(cor, k=-1) # below main lower triangle of an array\n",
"cor = cor.stack()\n",
"cor[(cor > 0.55) | (cor < -0.55)]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We will drop the non-informative columns found above. Note that we missed a few columns in our munging process above such as `funded_amnt_inv` which is the amount funded by customers at that point in time which is post interest rate determination. We also missed `installment` which is the monthly payment owed by the borrower if the loan originates."
]
},
{
"cell_type": "code",
"execution_count": 132,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"df.drop(['zip_code','funded_amnt','funded_amnt_inv', 'installment', 'mths_since_last_delinq', 'total_acc'], axis=1, inplace=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Feature Space and Labels\n",
"Before we go on any further and verify what we already intuitively know. Let's make a quick (wasn't really quick) plot to examine the relationship between interest rate and FICO score. We also can finish this post without replicating the same image from original post from long ago. Arguably it was a lot more intuitive to do in native ggplot back then."
]
},
{
"cell_type": "code",
"execution_count": 138,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAFaCAYAAADRpx5PAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x11145bf10>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"plot_df = df.query('last_meanfico > 600 & int_rate <28')[:3000]\n",
"\n",
"sns.set(font_scale=1.2, rc={\"lines.linewidth\": 1.5}) \n",
"\n",
"\n",
"\n",
"g = sns.lmplot(\"int_rate\", \"last_meanfico\", x_jitter= .7, y_jitter= .1,\n",
" data=plot_df, hue='term',lowess=True, size=5,aspect=1.4, legend_out=False,\n",
" scatter_kws={ 's':20, 'alpha':.6})\n",
"\n",
"g.set(xlim=(2.5, 28),ylim=(580, 880),alpha = .5)\n",
"\n",
"g.savefig('1.png',transparent=True)"
]
},
{
"cell_type": "code",
"execution_count": 139,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"y = df.int_rate.values\n",
"df.drop('int_rate',axis = 1, inplace=True)"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"(array([ 6. , 6.03, 6.49, 6.62, 6.99, 7.12, 7.49, 7.62,\n",
" 7.69, 7.9 , 8.19, 8.39, 8.67, 8.9 , 9.17, 9.49,\n",
" 9.67, 10.15, 10.49, 10.99, 11.44, 11.67, 11.99, 12.39,\n",
" 12.49, 12.85, 12.99, 13.35, 13.53, 13.65, 13.66, 13.98,\n",
" 14.16, 14.31, 14.47, 14.49, 14.64, 14.98, 14.99, 15.31,\n",
" 15.59, 15.61, 15.99, 16.24, 16.29, 16.49, 16.59, 16.99,\n",
" 17.14, 17.57, 17.86, 18.24, 18.25, 18.54, 18.92, 18.99,\n",
" 19.2 , 19.22, 19.24, 19.47, 19.52, 19.97, 19.99, 20.2 ,\n",
" 20.49, 20.5 , 20.99, 21.18, 21.48, 21.99, 22.15, 22.4 ,\n",
" 22.45, 22.9 , 22.99, 23.4 , 23.43, 23.7 , 23.99, 24.08,\n",
" 24.5 , 24.99, 25.57, 25.8 , 25.83, 25.89, 25.99, 26.06]),\n",
" <matplotlib.axes._subplots.AxesSubplot at 0x10b435bd0>)"
]
},
"execution_count": 38,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAi0AAAFyCAYAAAA52erNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmZnVwt8s9ffjwMQ3NgAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x10b42b6d0>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"np.unique(y), pd.Series(y).plot(kind='hist',alpha=.7, bins=20, title='Interest Rate Distribution')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We're about to blow up the feature space by dummy-encoding our categorical variables! This will make our dataframe 8x bgger in memory going from 60 to 470MB."
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"Int64Index: 235629 entries, 0 to 235628\n",
"Columns: 209 entries, loan_amnt to last_fico_range_845-850\n",
"dtypes: float64(200), int64(9)\n",
"memory usage: 377.5 MB\n"
]
}
],
"source": [
"df = pd.get_dummies(df)\n",
"df.info()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Gradient Boosted Regression Trees\n",
"\n",
"Like the original post we will use GBTs. The tradeofs are listed below:\n",
"\n",
"Advantages\n",
"\n",
"- **Heterogeneous data** - Deatures measured on different scale such as employment length vs annual income\n",
"- **Supports different loss functions** - We can pick a loss function that suites our problem. If our data contains a lot of missing labels like huber loss, or something specific for ranking problems.\n",
"- **Automatically detects (non-linear) feature interactions**,\n",
"\n",
"Disadvantages\n",
"\n",
"- **Requires careful tuning** - RF are faster to tune, they have essentially one parameter\n",
"- **Slow to train** - But fast to predict\n",
"- **Cannot extrapolate** - It is not possible to predict beyond the minimum and maximum limits of the response variable in the training data."
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"from sklearn import ensemble\n",
"from sklearn import datasets\n",
"from sklearn.utils import shuffle\n",
"from sklearn.metrics import mean_squared_error\n",
"from matplotlib import pyplot as plt"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"X, y = shuffle(df.values, y, random_state=30)\n",
"X = X.astype(np.float32)"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"offset = int(X.shape[0] * 0.75)\n",
"X_train, y_train = X[:offset], y[:offset]\n",
"X_test, y_test = X[offset:], y[offset:]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Training & Tuning\n",
"We won't spend too much time in this post tuning, but in general GBTs give us threes knobs we can tune for overfitting: (1) Tree Structure, (2) Shrinkage, (3) Stochastic Gradient Boosting. In the interest of time we'll do a simple grid search amongst a hand chosen set of hyper-parameters.\n",
"\n",
"One of the most effective paramerters to tune for when working with a large feature set is `max_features` as it introduces a notion of randomization similar to Random Forests. Playing with max features allows us to perform subsampling of our feature space before finding the best split node. a `max_features` setting of .20 for example would grow each tree on 20% of the featureset. Conversely the `subsample` feature would use 20% of the training data (all features). Subsample interacts with the parameter n_estimators. Choosing subsample < 1.0 leads to a reduction of variance and an increase in bias.\n",
"\n",
"### Loss Function\n",
"\n",
"- Squared loss minimizes expectation. Pick it when learning expected return on a stock.\n",
"\n",
"- Logistic loss minimizes probability. Pick it when learning probability of click on advertisement.\n",
"\n",
"- Hinge loss minimizes the 0,1 / yes-no question. Pick it when you want a hard prediction.\n",
"\n",
"- Quantile loss minimizes the median. Pick it when you want to predict house prices.\n",
"\n",
"- Ensembling multiple loss functions\n",
"\n",
"Adapted from: https://github.com/JohnLangford/vowpal_wabbit/wiki/Loss-functions"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"from sklearn.grid_search import GridSearchCV\n",
"\n",
"param_grid = {'learning_rate': [0.1, 0.05, 0.02, 0.01],\n",
" 'max_depth': [4, 6],\n",
" 'min_samples_leaf': [3, 5, 9, 17],\n",
" 'max_features': [1.0, 0.3, 0.1]\n",
" }\n",
"# param_grid = {'learning_rate': [0.1],\n",
"# 'max_depth': [4],\n",
"# 'min_samples_leaf': [3],\n",
"# 'max_features': [1.0],\n",
"# }\n",
"\n",
"est = GridSearchCV(ensemble.GradientBoostingRegressor(n_estimators=100),\n",
" param_grid, n_jobs=4, refit=True)\n",
"\n",
"est.fit(X_train, y_train)\n",
"\n",
"best_params = est.best_params_"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"CPU times: user 2h 15min 34s, sys: 5min 43s, total: 2h 21min 18s\n",
"Wall time: 3h 28min 36s\n"
]
}
],
"source": [
"%%time\n",
"est = ensemble.GradientBoostingRegressor(n_estimators=2000).fit(X_train, y_train)"
]
},
{
"cell_type": "code",
"execution_count": 48,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
".7644790837986264\n"
]
}
],
"source": [
"est.score(X_test,y_test)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Examing results\n",
"In terms of simple model accuracy we know we're doing pretty good with the above score over 3000 iterations gets us in the 90s, 1000 gets us in the 60s.Let's examine how drastically accuracy increased with each iteration and see which model features were most powerful in predicting the issued iterest rate."
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA0IAAAGJCAYAAACn7oMqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nICQiIiIiIg3n/wMz6UxEvqiJPAAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x10c56fd90>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"sns.set(font_scale=1, rc={\"lines.linewidth\":1.2}) \n",
"\n",
"\n",
"Iterations = 2000\n",
"# compute test set deviance\n",
"test_score = np.zeros((Iterations,), dtype=np.float64)\n",
"\n",
"for i, y_pred in enumerate(est.staged_predict(X_test)):\n",
" test_score[i] = est.loss_(y_test, y_pred)\n",
"\n",
"plt.figure(figsize=(14, 6)).subplots_adjust(wspace=.3)\n",
"\n",
"plt.subplot(1, 2, 1)\n",
"plt.title('Deviance over iterations')\n",
"plt.plot(np.arange(Iterations) + 1, est.train_score_, 'dodgerblue',\n",
" label='Training Set Deviance', alpha=.6)\n",
"plt.plot(np.arange(Iterations) + 1, test_score, 'firebrick',\n",
" label='Test Set Deviance', alpha=.6)\n",
"plt.legend(loc='upper right')\n",
"plt.xlabel('Boosting Iterations')\n",
"plt.ylabel('Deviance')\n",
"\n",
"\n",
"plt.subplot(1, 2, 2,)\n",
"# Top Ten\n",
"feature_importance = est.feature_importances_\n",
"feature_importance = 100.0 * (feature_importance / feature_importance.max())\n",
"\n",
"indices = np.argsort(feature_importance)[-10:]\n",
"plt.barh(np.arange(10), feature_importance[indices],color='dodgerblue',alpha=.4)\n",
"plt.yticks(np.arange(10 + 0.25), np.array(df.columns)[indices])\n",
"_ = plt.xlabel('Relative importance'), plt.title('Top Ten Important Variables')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Partial Dependence Plots\n",
"When working with tree models such as GBTs or Random Forests, I like to take a look at Partial Dependence Plots to understand the functional relations between predictors and an outcome. Tese plots capture marginal effect of a given variable or variables on the target function, in this case interest rate.\n",
"\n",
"In the first plot, tt is interesting to note that below $25k annual income, Debt-to-Income ratio has a large effect on the interest rate. The margin gets significantly larger after 75k. The steepness of meanfico is especially interesting, it looks like no real distinction is made for customers with a score over 760."
]
},
{
"cell_type": "code",
"execution_count": 55,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA48AAAJXCAYAAAAzc2H5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd8VFX6+PHP9GTSeyWElhuQ0It0ULpgQ0RFEcXe+Lru\nunZs23Utu5Zd14b1JwKyCrKAUkSKCIQECJdQE0jPpJdp9/7+mDAQkpBCkknIeb9eecHcNs9MkpN5\n7jnnORpVVREEQRAEQRAEQRCEC9F6OgBBEARBEARBEASh4xPJoyAIgiAIgiAIgtAokTwKgiAIgiAI\ngiAIjRLJoyAIgiAIgiAIgtAokTwKgiAIgiAIgiAIjdJ7OoDGXHXV7HYpB3vgwH6mTJmC2ezTrPP0\nei0Oh4Kqquzfn0JeXh533nlXm8S4bdtW5s27GUVx8vbbb6GqKoqicP/9D2I0GpsZtw6Hw9kmcdZH\nUZz8/PPPLFhwe4uv8eabb5CYKDF16vRWjKz9nPlZaW/p6TKlpaXMnDmrRed36xajaeWQhHrk55d1\n2NLXQUFmiooqPR1GvTpqbB01Lui4sW3M+IGHN95L6oJ0NBrPNDthYX6ivWsHHaW966i/CxciYm4f\nXSHmlrZ3HT55bA9Op5Py8rImJ46ZmZn4+voQFBSMRqMhJyeLyspKrFYrCxfe2SYxHjlyGKvVCoBW\nq+Ohhx5BVVVsNluzE0eA9v67vGvXLnx8mpeYn6+ysoLBgwe3UkTtrz0+DCmKw/3/r7/+mqKiIpxO\nJ88993ybP7dw6dLrdZ4OoUEdNbaOGhd03NiyK7LIq8gjo+wk3f3jPR2O0AV01N+FCxExtw8R8wWe\np12epYOrqqrCy8u7yccnJ+9FURSuuOIKgoODAJg06cq2Cg+AvLw8Jk68otY2jUaDyWRq0+dtLVlZ\np7n33vsv6hpPPPEUBoMOu739ekw7o6ioGAAefnixhyMRBEFousLqQgBS81NE8igIgtBBiTmPNZrS\nKeR0upIWX19fqqoq2bt3r7v3TxAEQRCEliuqtgCwv2CfhyMRBEEQGiKSx2bQ6VzdwUFBQcycOQur\n1conn3xCWloaqtohhu8LgiAIQqdkqXL1PKbkNz15PFqczpt7XmurkARBEITziOSxEXa7nS1bNrFj\nx3YyMzMBUBSF0tIStFotVqsVLy8vj03uF4RzffTRR54OQRAEoUUsVgtSiERqQUqTz/nvkW94O/mN\nNoxKEARBOJdIHi/Aaq3mm29W4u1tRqfTIcuHsNls6HR6Nm/eRFBQEHPnzuXkyZMcP37M0+EKAna7\n3dMhCIIgtEhRtYUpPaeQW5lDbmVuk85Jzt+LpdpCtaO6jaMTBEEQQBTMuaCSklL8/HwZPnwEACtX\nrqCoyEJkZCRarYakpAEYDDpGjx5Njx49PRyt0FyqqmK1VqOqYDKZ0Grr3kvJy8vD4XBgNBoJDQ0F\nXAWWMjJOAhqCgoIIDw9v58gFQRAuPZaqQobHDMd/XwD78/cR0X1qo+fsy9sLuCq19ggQf4cFQRDa\nWpslj5Ik6YD3gARABe4DrMBHgALsBx6UZdljkwUdDgcHDuynurqK4ODgOvu9vb1RFBWr1VqTXGgo\nKyvF19eX0NAQcnKyUFUVX1/fVo+tsrKS9HQZu92O3W4nIyOD2bOvbvXnaQ/796dSWFiIl5eXp0Nx\nUxSFNWtW88UXX+Dn58v06TO49trrah2TmprKa6+9io+PD97eZp555llMJhP//ve/OHr0KAaDgYCA\nAB5//PcefW2qqpKSkszmzZvp27evx+IQBEG4GEVWC6HmUJJCB5BakMKVjSSPuZW5ZFWcBiC7XCSP\ngiAI7aEtex5nAYosy2MlSZoA/LFm+1OyLG+RJOkd4BrgmzaMoUG//voLqgr+/n7MmXNDnf1OpxM/\nPz+mT59BXl4OVVVVmEwmJk26kv3791NUZOHyy0fh62tulaUj1q1bS2VlJRqNBkVR8PLyIjIykssu\n649Op+OGG26st2eso1IUJ5s2bSQzM5PY2G689NIfOlT8JSUlfPDB+6xY8Q2VlZXcf/99XHnlZPz8\n/ACwWq38+OMPPPLIYgYNGsw777zNmjWrmTnzKg4cOMC///0eAE899SS7dv3CuHHjWyUuRXFQXV3N\njz/+WGef0Whk8uTJdbYXFBSQn5/PCy+8hF4vBhMIgtD5KKqCpdqVPPYPG8CPGRtYeNkiAr2CGjwn\nJW8vgaZAAkyBZFdktWO0giAIXVebfdKUZXmVJEnf1TyMB4qAybIsb6nZ9j0wFQ8kj2lpBzEaTVx/\n/Zxa2/fv3w+o9O+fhE6nQ1VVd8IzdOhw8vLy+emnLaSmpnDDDXNbbY3FEyeO4XQ6L5l1+VJS9rFz\n5w5Gjx7DokV3ezqcep06dco91Fir1TJo0CB+/XUXkya51tI0mUwsXvx/AMjyIY4dO8oNN9yIv7+/\nO3HMycmmoqKCmJjYZj233W6noqIcAB8fXwwGA4ricO+Pi4tn8uS6d9y1Wq17DcdzRUXFkJQ0sFkx\nCIIgdCSl1hIUVSHEO4RbEm9jY8YGhn6axOuT3mJ2r2vqPSc5fy8DwwZjV+xkieRREAShXbRpN4Us\ny05Jkj7G1cM4F5hyzu5yIKCxaxw8uJ/u3bu7kzhFUTh+/DgOhwOHw9HgEhkajQadTofT6axzjEaj\nYf78+XV6wsrLSzl06BAOh52oqCj3eWPHjiUzM5NfftnJkCFDWLRoESaTCUVxYLU6UBSl1pfD4cBq\ntZKTk0N2dra7R7Ehqqry4IMPYjDoGjzG4XBQWlqK1WpFp9Oh0+nc17TZbFRXV2O323E6nTgcDsrK\nyrDb7RQXF1NaWoqiKOh0OrRaLRqNBo1Gg6qqteI+N54z+8/99/yYzz/mzHsbFRXFX//611rn2Gw2\nCgsLqaiowGazYbPZ6NWrV70J+Pr166mqqnI/diVNUQwbNgzQoaq1e3qXLVtGdXU11dXVLFq0qNEe\nTtd7WYzZ7I2qOtFqwWg0UFlZUefalZWVPP/8Ery9vQkNDXbvz87O5tNPPyU+vjs9enSvc975Vq9e\nzenTp92vx9/fH4AJEyYQERGBRqOhW7du7uN7927+8CtFUaiqqnKvR3qGyWSq933OzMyksLDQfW51\ndTVVVVUkJCQQG9u8hFgQBOFiWKpdbVGoORT/kHA2z9vBU1t/x9IDHzSYPO7L28ug8CGcKsskp1wk\nj4IgCO2hzce4ybJ8uyRJEcAvwLkTw/yA4sbOHzp0GLt2/VIrORk1ajRmsw9ms5m9e3fXSs5UVcVo\nNDJixOVYrVZOncogMzOzTnKUmppaa56j1WolKyuL8eMnkJqawqBBQzhx4hgWi4WvvvoKu91OfHw8\n3t7erF+/3j088EwidiZZ1Wq16HQ6DAYDJpOJG264kRMnjp9XjdUVq9PpqEn4FL744stzki215piz\n/+p0Ovz8/DGZjDgcThTFiestUTEYjJhMJgwGA3q9Hp1OR0KChNFoIjQ0lKCgIHQ6HQ6HoyaxUNHp\ntDidrp7Vc7/OdW5v2LlWrFhBcXGxOzE/kzzeeOONBAYGAq5k8Ywvv/ySqirXvFJfX1dPm9FoJCIi\nEi8v7zrXnzJlKk7n2e+X3W4nNXUflZWVeHt7ExkZXev4O+5YhEajwdvbu9HE8U9/+iN79uzm8OHD\nDBs2nMjIaBRFwelU6NGjJ5GR0Zw6lcn69euxWCwYDHpef/0NFEXh979/nPXrN5CVleX+mYiJieWL\nL76sdYNi8uTJDBo0qNbz3n33PfzwwwaSk/ehqqr79f3440auvPJKBg0aVOs1A2zYsJ59++qWrJ88\neTIDB9btadywYT3p6UfQ62vfhBgzZiz9+vWrc3xxcQm5ua6KhhqNFi8vL6KiYggNDcPPr9H7OoIg\nCK3GUm1Br9Xjb/KnoKwcnVbHpG6T+e+RlQ3ewEzO38tNibfiVJ0cLxEVzwVBENpDWxbMuRWIlWX5\nz0AV4AR+lSRpgizLm4EZwA+NXScmJrbBYYGHDx+isrKSOXPmoqqq+w/M99+vZt++vQwePJQTJ06w\nYMFCjEZjozGHhoaRmNgXb29vPvvsE8aOHcsjj/wfGzasJyIigqSkAXXOMRh0jc55TE5O5re//S0A\nqoo70TAaDej1hnabp2YwGDAYDDX/v3Dcv/yyA4DZs+ve8b3nnvvc12uKRx99rFlxRkZG1dkWHx9f\n85x1466vYJFWq+HYsWOsXLmy1nYvLy9uuWU+V199NVOmTKaoyILT6SQlJYWbbrqJv/3tL9jtDn77\n29/x8ssvMX78eOLiurN37x4GDRqEwWDkgQfuJzGxL3/845/Jy8sjLi6OsLCwRl/XTTfdxPXX151f\n25A5c+YyZ87cZh3fHKNGjW7W8YIgCG3FUl1IkCm4VpKYFDqAwupCsiuyiPZ1DdmvsFdw1/8WoKKS\nV5nLoPDB5FRkse30T54KXRAEoUtpy6xlBfChJEmbAQOwGDgEvCdJkhE4CHx9MU+Qk5NDREQEZrO5\n1vYZM67i00+XAq7CN01JHAH69bsMgMsvH8XWrT9x+HA6NpuNwYOHEBIS0uI4NRrw9fVr8fntzWq1\nsnTpUh555JF69zc1afQUrdb14aNnz57cfvvtrFq1qtb+4uJiVq9ezZIlz7NgwQK0Wg133nknAwcO\nxGw28+KLLxIREcETTzzJo4/+H1999RVms5m///010tMPU1paBsDvf/84JpOR229fyMSJk9r9dQqC\nIFwqLNUWgr1qVz2P9o0h2CuYlPx97uTxy0OfcbDwADcnzmdoxHBifGOJ8o0huyLbE2ELgiB0OW1Z\nMKcSmFfPromt9Rxjx45nzZrvyMzMrLXdarW6C46Ehoby7ber6u1Bu5C7776Xd955i2PHjpKYeHHL\nHxiNRv7wh5dxDUOFM8NWBw4cwKxZHW/5DZPJxOuvv8GKFcvZvn07jz/+hKdDahZFOTuE1N8/gOuu\nu67OMVqtDn9/fyZMmFjrvG7d4pgxYwaLFz/C1VfP5quvltU6LyQkhN2797RZ7IIgCF2RpdpCsHft\nm7QajYb+oQNJLdjH9B4zcSpO3t33T+4b+BD3D3rIfVyUTxS5lTk4FAd6rag4LQiC0JY6fCt7prhK\nQzQaDTabDY1G455vqNfr3b2NrkqWSoPnN8Tf35/4+PiLThyBmsI99prhOK7ExjU3s3WqtbYFvd7A\nqFGjWLt2radDuSh6vZ7AwIZLvZ/PaDQycuQIUlNT8PFp/fU7BUEQhLqKqi0Emequt3xmzUeANce/\nw1Jt4dZ+C2odE+0bg6Iq5FXmunsoBUEQhLbR4ZPHPXt288svO5t07NSp05g2bTorVnzN/v0pJCUN\nJD8/nzvuWNTs5921aycJCQnNPq8+VquNZ599rlWu1Z6io6N5+ulnPR1GuyoqKuKjjz7mjTf+0eGH\n5wodnyRJBuADoDtgAl6WZflbz0YlCB2PpbqQEO+600MGhA1k1ZEVKKrCP/b8nQX97sDP6F/rmDDv\ncHQaXa25kUL7Em2dIHQdHT55HDJkKP37J9W7z+l0sHPnDm688SYA93IETqeT6OizRXYutExGQ3x8\nfKioqGhBxHU1VgG0ozpTRbYrCQwM5MUXXxSJo9Ba5gP5sizfJklSEJAMiA9UQpdSX7XU81mqLfQK\n6F1ne1LoQE6VZ/LHHS+SUXaSTwc9XOcYnVZHpE8UWeVZDI1otbCF5hFtnSB0ER0+qzEajfj6+tb7\nFRAQiFarJSAggICAALy8XCuBWK1WvL1dS0CcWaKiuWJj4ygrK2ulV1H/WpRCx9MVE2ahTS0Dzgw7\n0ALNb4wEoRNbJn/J5Z8PJqv89AWPs1QX1pnzCNAzsBdmvQ9v7v07r0x4k3BzeL3nR/pEkVMh1nr0\nINHWCUIX0eF7Hlui7l3O5idv5y58LwiC0BKyLFcASJLkh+vD1dOejUjoyDZnbmRc7AS0mg5/X7fJ\nlh3+ktyKXOZ9ex2rrvueYK/6K5cX1VNtFUCr0TIiaiTRPjHM6tVwgblo3xiyykXy6CmirROEruOS\nTB7tdrt7CKter6e62orB0LTlOs7w9fVttWGrZ6qrCoLQ9UiS1A3X0kVvybL8ZUPHBQWZ0et17RdY\nM4WFddzlhjpqbM2Jq9Raytxvr2HX3bsYFj2sDaNyaY33bGXaSvxMfkzuOdm9rcJWwc3Lb+b9q9/H\nqDPy8+mfWH3Lap7d+Cx/3vMCH17zYb3XKrJa6BERW29sa277Di+91wVHhfQKjSe3IrfD/ix0BU1t\n66BjtXed8WdGxNw+RMz1uySTR1VV3fMMDQYDNpsVaN6bqdVqW7HnsXP2YB48eBA/vwB8fUXVUUFo\nCUmSIoB1wAOyLG+80LFFRZXtE1QLhIX5kZ/fWsP4W1dHja25cZ0oOQHAgcx0uhukNorKpTXes8Kq\nQhasvB2zwcyO+XvxNbj+TuzM3sG3h7/lxR/+yIDQgQSYAknyHc6Nfebzn5R3631eVVUprCpEa3VN\nPanvmHLKLxjPjNhryK7IvujX1Rk/LHYEzWnroOO0dx21/bgQEXP76Aoxt7S9u3TGxnRonbPn8dCh\nQ5SXX/gP9qWooqKyFXudhS7uKSAAeE6SpI01X16eDkroeIqqLQBkd5J5e6/v/hs9AnriY/Dh7b1v\nurenFR7ApDPxQep7fH7oE6bFz0Cn1ZEY3JcjxenYnfY61yqzleJQHPUOW22qwRFDmdlzVovPFy6a\naOsEoYu4JHsez9fSHsSuXjjFYDC0qNhQZ3fw4EHKy8u49trrPB2K0MnJsrwYWOzpOISOr8haBEB2\nJ5i3d7L0BB/sf4/Pr/qaMlsZD/1wD7dfdicRPpGkWQ4wq+c1HLKksSnzRz6d+f8AkIISsSt2jpUc\nRQpOrHW93bm/AhDqHdbur0VoHaKtE4Suo9P3PDaUGCqKAriW7dDpmp8jO53OLj9stbi4GB8fH0+H\n0e5OnjxBaGiop8MQBKEL6Uw9j/9JeZcRkZczodskruo5mz5BEp+lLQXgYOEB+oX254mRzxBhjmRc\n7EQA/E0BRPvEIFvSal1rf0Eqd69byG+G/o4AU2B7vxRBEAShmTp18mi1Vte7Hl9gYCCHDh0EoKqq\nioCAgGZfe//+FLp163bRMXZWDoeDkpISgoKCPB1Ku6qurubQoUOMHTvO06EIgtCFdJbkUVVVVh/7\nlrmSa31ljUbD1Pjp/HRqM6qqklZ4kH7B/ZgWP4M9tx3AW+/tPjcxpC9ploPuxydLT3DTd9dzbe85\n/H7EM+3+WgRBEITm69TJY3W11b2247mGDh2OzWZzPz5TPKc5ysrKGDFi5EXFd1bnG/5qtVq7XOII\nUFJSQnx8d0+HIQhCF+MettrBksdye+1576kF+8iqOM20+JnubeNjJ7ErZydHi49Qaiuhb8hlABh0\ntW/uSkF9kS2HAMivzOfGb69leORI/jL+1S4/TUQQBKGz6NTJo9FoxGKxkJGRwTffrGDFiq9ZtWol\nP/yw3p1U6nQ6SktLm33twMBAVq/+rlXitNvtHDt2jCNH0jl0KI19+/Zx4MCBWgluR2MyGRkzZgzZ\n2Rde2PlS4+vrS0pKKikp+zwdiiAIXUhRtYVonxiyy7PqnTLx/LZn+OfeN9o1ppOlJ+j7QQ++PbrK\nvW31sf8yKmoMId5n12scEj4Ug87IRwf+Q4ApkCif6Hqv1zekH7IlDVVVWbj2FqJ9Ynhn8n/QaTvG\nkg2CIAhC4zp1wRxvb28SEhLYvXsXUVEx6HQ6HA4HcXHxBAe7qraNGTOOL7/8nHvuuY+VK5ej1+uZ\nPfuaRq/dv/8AcnNz2bjxByZNuvKi4pw4cSKbN29Cq9Wi0+kwGo04HHb++99VVFSU8+yzS9zrUnYU\ner2Bnj17oyiOWglkTk4Oq1atwmaz0a9fP269dYEHo2x9Pj4+/OlPfwZUSkqKASgvL8fhcKCqKh9+\n+CFeXl7u9al0Oj0BAf7MnHkVUVH1f2ASBEFojKXaQr+Qy9iQsY5SW0md+X9rj69mUPjgdo3p04Mf\no9Po+P2WRxkVPYZQ71BWH/uWhZctqnWcQWdgTPRYPkv7hKTQAQ32IkpBiRwrOcp3x1axvyCFXbem\n4qUXBTkFQRA6k06dPJ6h1WrRajVoNBrO/5ulqqr7D5mXlxc6XdPvcJ67XuTFct1JVt3/V1VXUR9X\n7J2nA1ij0dTEr+Lnd2muh6XRaLDb7fz9739HVRV8ff0wGo0AxMbGMGvWbJxOJwAOh52SklJCQkSB\nHUEQWq7YWkT/kAFsyFhHVnlWreSxoKqAYyVHifaNabd4bE4bn6d9wp/GvcKnaR/zwIa7GBc7kcNF\nMjN61F0SY3zsRNadXEvfkH4NXrNPsIRTdfLElt9y+2WLCDeHt+VLEARBENpAp08eDxw4wOzZV5Oc\nvBdFUdDr9WRmZjJ9+kwMBgO7du3kuuvmADBjxlVNvq7VauX06dPccceixg9uxKZNG5k7dx52u42K\nikrKykrx8vJi3rx56PV1C/50FCdPHqeqqooJEya5t0VFxTBo0BAPRtX2tFoNJpOJF154sVk3GwRB\nEFqqqNpClG80gaZAsiuyaiVhv+b8AkBeZW67xbP2+Grsio1r+8xhZNTlPLrpYb49spLb+i0kxi+2\nzvFnqqr2Db6swWv6GnyJ848nryKHBweLVR0EQRA6o06fPPbq1Ysff/yh1jZvb293FdaAgAC2bt3C\ntdde36zrmkwm/P39sdls7l6nlho58nLWrl2DTqdHr9dhNJqw221s2bIFs9mbxx77XYcsFhAbG8fK\nlctZv349wcHBKIpCeXk548aNo18/1webqKj2uxPeXhRFRavVUF5eBsCRI0dYv349AAaDHqPRiE6n\np3v3OHdV1oAAUWL+UiFJ0nygH/BHYI4sy0s9HJLQBViqLQR7BRPlE11nrcddOTvxNfi1W/Koqiof\n7H+PedIteOu96RnYm1XXfn/BcxKD+zI5bipjY8Zf8Lih4UOJ9o0lwhzRmiELLSTaO0EQmqvTJ4+J\nif1wOp0UFBS4iwycGVZ5JiFr6XqNZ+YmXmzymJjYl8TEvrW2KYqTHTt2kJy8t1asHYlOp+Pqq6/l\n7bf/SU5ODhqNBkmS3InjpUxRXD8zWq0Go9FIQUEBoGK3O7DZrDidCqWlpe7ksaSkWCSQlwBJkv4C\nxAJDgL8Cd0qSNEiW5d94NjLhUldsLSbQFESUbzRZFbULle3K2cmU7lNZeWQ5VqcVk65t58i/nfwP\n9hek8uYV7zT5HI1Gw+ezvm70uHemvH8xoQmtSLR3giC0RKdOHlVVZcOGdURFRXHbbQswGFxJ3qpV\nK9m1aycjRlxOdnY2d955V7OvfeLEMaqrqzGbfS46zhUrvqa6uhrXkh2upMTpVEhK6s9f//pKh0wc\nAYqKinjvvX8xf/58hgwZ5t5+poBOR+11VFUVi8VCcXEx8fHx9Q49PX78mDtBPFd8fDwGQ+1fi6Sk\nAbzyyqttFq/QoUzD9UFqtyzLJZIkTQZSAfFh6hKmqAql1hICvTyzPJFDcVBiLXb3POZUZLv32Zw2\nkvP2cNeV/2LlkeXkV+YR69d2axD/dGozf9j5PB9N/4w4/9Zftkir6Txz/LuATtXe7cjeTkJQAsFe\nIY0fLAhCm+nUyeP27T8THR1dp3rqNddcx9KlH7Ft21bMZnOz563Jchqpqak8/PAjFx2jojgpLy9n\nyZIXLvpa7UlRHKSlHeD551+sUxinIyaN6enpbN68kcJCCwaDgZCQEAIDA+nePQ6ttu6PeVpaGg6H\no8727t3jADEMtQtznvfYVM824RKz9vgalmx7il23pnjk+YutrsrOQTXJ49683e59qQX7cKgOrug+\nBa1GS15lbpslj6qq8viWR3lw0GKmxs9ok+cQOpRO096pqsqSn5/kQMF+ZvSYxfx+CxgfO1HcjBAE\nD+i0yaOqqpSVlXHzzfPr3X/bbbfz29/+hr/85W/Nvva+fft49NHHWqUK6qlTp4mL61yLziuKK6k6\nU2ioo9Fq6/bUFhTks3DhHYSFNa163/z5t7Z2WMKlYRnwJRAsSdKjwG3AF54NSWhrBwv3c7L0BCXW\n4jpLZLSH4uoiAIK8XMNW1xw/u8bwz6e30j8kCV+DL6HeYeRX5bdZHDuzt5NRepL7Bj7UZs8hdCid\npr3TaDSsnbOR7Vk/81naUm7//mZCvcO4KXE+8/suaNdKxILQ1bVZ8ihJkgH4AOiO627Wy8Ap4Dvg\ncM1h78iy/FVLrl9WVnbBpSI0Gg3PPbcEvb55L1FRFLy8vFpt+Yzc3ByGDOk81UnPJI7t3buoKErN\nUiuND+GtrKwkMjKq1rEzZza9kq4gNESW5T9LkjQdyAC6Ac/JsvxdI6cJndzR4iMAyBaZEVEj6+x/\nL+Udruszl1DvtlmSx1JtQa/V42vwI9onmuxy19QAu9POxwfe594BDwAQbo5oVtGcJ7Y8RrWjmpfH\n/QVfg2+jx3+WtpQZPWYR4i2GBXYFna2902g0jI4Zy+iYsfzJ+jdWpH/NZ2lLeSf5n7w/bSmT4i5u\nTW5BEJqmLfv75wP5siyPB6YDb+EaW/+qLMuTar4aTRxtNhvl5eV1vlJSkhk2bDglJSXuL9e8wrNa\nMvRQq9Uybty4Zp/XEIulkIgIz1aVUxSHOyls7Dho/8RRVVWWLHmWTz5ZilarqfP166+7ah3/1ltv\ntWt8QtchSVIMcIUsy78F/gnMkyRJlIW8xB0rcSWPhywH6+wrs5XyzNYn+DztkzZ7/iKrhSBTMBqN\nhl6BfSixlfBjxnq+ObKcclsZ8/vdDkC4ObzJyeORonQ+OvA+27K2Mm3ZRF7avoQ///Iy5fbyeo8v\ns5Xy7dFvuKXvba32uoSOrTO3d/6mABb2X8S6GzaxeMhvuHXNjW36OyoIwlltOWx1GXCm9JoWsAND\nAUmSpGuAdOD/ZFmu/y9ZjT17dvPLLzsb2Len1uOpU6cxbdr0iwq6uNjSqgVsioqKCA0Na7XrNVd1\ndRUWSyGRkZHu5DAjIwO73U5ycjIFBQVoNBoURWHWrFkMH173rntbW7XqG2bNms306fXPsdmxYzv9\n+/fHy8uJC/TUAAAgAElEQVTbva2jFhkSOr3PcA3jAjgNbAE+AaZ6LCKhTamqyrGSYwR7BSNb0urs\n35efjIrK6mOreGTIo20SQ1G1haCaYj3xAT14YsQz3Ld+EUFewdw14D58DK7Cbc3peXwr+Q0mxE7i\ng+mf8squP3Oi9Dg/ndpEvH8PHo6+z31cZlkG/+/Q5xwpPkywVwgTYidd4KrCJabTt3cajYbFQx+j\nm38cj/xwPydLj/PEiGfFZwRBaENtljzKslwBIEmSH65E8mnAC3hPluW9kiQ9BSwBfneh6wwZMpT+\n/ZNqbXM6HWzd+hPz59e+Q2oytU758qSkga1yHQCHw9HsobOtacOG9fTq1YvBg4cC4HQ6OXz4MFlZ\nWdxww41ERkZ6LDaA4uJitmzZzAcffNTgMb169UaWDzNwoOv7oqpKO0UndEHBsiy/CyDLshV4T5Kk\nBzwcU5vZmb2DgWGD8NJ7eTqUVnPYIvPC9mf47KplTTq+sLqQEmsxNyfeyqGiQ3X278ndTYxvLHvz\n9nCqLLPRYjV5lXmENLMaZFF1EUFewe7Hi4c8xr78ZDZmbGBR0j3u7WHe4RwrOdrgdVRVpcxWSrG1\nmK/kL/hq9jf4GHxYMvolAJ766XesOf4dD4+7z338Axvupri6iDj/7jx9+RJ02uYVmBM6tUumvbu+\nz1yifKK5/fubySjN4PUr3mrzJW0Eoatq06xGkqRuwArgLVmWv5QkKUCW5ZKa3d8AbzZ2DbPZCx8f\n71rbVFXFx8eHzZs3MmdOw0Vd8vPzCQtreq+foiikpaWxbt06fH198fLyonfv3owaNeqCianB0PAf\nWx8fHyyWAo8NXR05ciT/+c9/2LRpEwB6vZ6bbroJSZIAhZycs4tRf/rpp1RUVGA2m7n77rub9d61\nVEhIENdccw3vv/8e8+bNc28/duwYy5cvR6PRUFFRwZQpk93vs1ar5dChg4SHh3s8+W0uo7HT1qjq\nKqokSZopy/IagJrS9RccHdFUkiSNBP4sy3KH6Nr5Jn0596y/gydHPMujwy54D69T2Zu3m/Un/0dB\nVUGT5igeLT6CQWtgcvepPPlT3fdhb95uZvW6hs2ZP7L62H+5d+CDF7zedd/MZF7iLbw0bUmTY3b1\nPJ5NHjUaDW9Pfo9TZZm1liUIN4ezI3ub+7FTcXKgMBVLtYUjRYdZevBDDtX0no6IvJzR0WNrPc/M\nnrO55bsbqLBVAPDt0W9IzU9hx/w9RPpENTle4ZLRJu2dp9q6UdFjWH39Bm5efQPzvr2Oj6Z/5rHl\ndwThUtaWBXMigHXAA7Isb6zZ/D9Jkh6WZXkXcCXwa2PXcTjq72WaOPEKVq1aic3mqHd4gqqqvPzy\nyzz33BL8/QOaFPPXXy/D4XDQvXsc/fpdRlVVFcnJyezZs5eHHnq43ucxGHTY7Q1Xth41agxffPEF\nDz74cJNiaG1hYRE8+eTTdbbXF/dtt7nm1Rw+fJh58+bx9dcrLliUqLVMmOD6+3Lu9zouLp5HH32s\n1nF2uxODQcfcufNYuHAhsbHdmDRpIrfeerYHuiMvsWE06rHZGp972tEYjUZPh9Ce7gU+kyTpzOSZ\nTOCiS/NKkvR4zXVaJRFtqX15e9mRvY1Q7zAe3fgQU7pP473Ud7h34IOYDWZPhnZRtp7ewuVRo9Fr\n9WRXuG6IHShIZUK3xj+7Hi85So+AnvQLuYy8ytw6iVxy3h6u7nUtvgZfvqsneXQqThRVwaAzUOWo\n4khxOu+lvMuzk59scvxF1iKCTLU/5HrrvekTlFBr27nDVv/yyx/4+MAHFFYV4Gf0J8IcwbzEW5gW\nPxO7Yic+oEedv1mXR43GS+/FuqPrGBo4hhe2P8viIb8RiWPX1ertnafbuj5BCXx//Q8s+H4e05df\nwdIZX5IQLHkiFEG4ZLVlN8hTQADwnCRJz9VsexR4TZIkO5AN3NPQyY1JSUkmMjKywXHtGo2GP/zh\nT80aypqdncWtt96KRqNxD5UdPnwETz/9JIqiNHu9SIDdu3dx/fU3NPu81qQoDtasWYPFYmHcuHH0\n6NHLvb0+6ekyy5evxNe38ep8rcFqtbJ27fdcd911dfadOnWKb775hgceOPuB7dtv/8uyZcsJCgpC\no9Hg4+PTLnEKlz5ZlpOByyRJCgHssiyXttKljwDX45pP5DF5lbmsTP+aNMtB7ky6h6dGPseITwfy\nxaFPWJR0rydDa7EqexVzVs3m/81eycRuV5BVU6n0QOH+JiWPR4uP0DOgF/H+PTHpTMiWQ1wePRqA\n3MpcTpefYlD4EBKCEvn7r38lrzKPcPPZJYFe2/03jhSn8+6U9zlafAQVFbti48v9XzIjum6bVp+i\nakuT1m4MN0eQX5nHseIjvPrrX3jzine4quds/Iz+TXoevVbP1PgZfJLyCa+X/wOD1sB9g8SyHF1V\nG7V3Hm/rwsxhrLxmDY9teoTpy6/g3Sn/EeuWCkIrass5j4uBxfXsGlvPtmbLy8vjzjvvuuAxzZ0D\nqdXqOHHiBNOnz8RqtaIoCsePH8PX1xdVVVsUZ0lJCdHR0S06tzUtWnQ3xcVFfPDB+2zbto0BAwZQ\nUlLCnDlz6xzb2Pva2rZt20p0dHS9vYYBAYEsW7as1tzR6upqunVrm0Wyha5NkqQhuG58BQMa1/Bu\nVFmWr7iY68qyvEKSpPiLj/DiTImfzpT46aiq6r7x9sCgh3lr75ss6HcnBp3BwxE23xGLK2E7WXoC\nwJ087i9IadL5x0qO0jOwNzqtjt6BCRyypLmTx+S8PQSZgoj37wFAhE8kP53axJyEG93np1kOsifX\nNYgmvUimm18c86RbeHX7q0y8ZjpeOq9Gi3cUVVsYENb4XPtwcwSVjko+S/uEIeFDuSmx/nWOL2Rm\nj9ksXHsLA8IGseq6tXjrvRs/SbgktUV711HaOi+9F/+88l+8s++fLFw7n98Pf5pHhvxGFNIRhFbQ\nqSdgtXYjcPPNt/D111/x/fdr0Wq1KIpCjx7x3HXXPR4tetNaAgOD+M1vfstPP21m3759HhtKe76U\nlFSefvqZBvf7+/tTVlZGUFBQzeO2H0ordFlLgXeBA8CZO0Ytu3PUTEFBZvT69i9W8n/jH+Llnc9z\nuDqFifETGzwuLKzj/N4VVRVh1BnxMfqw5aBr2eACRzZhYX7kWXMYEjUEufhgk2I+WX6M2X1nEhbm\nx8DoJDKqjrrPk1NTGRE7gvBwV8/e1N5T+KXgZ+4bs8h9flZlJqfLT4G5mlPW41wW0Y/fTXyUf735\nFt3/HUGAKYCjjxwlxNxwEZ0yZwndQqMajdfg5xo18tmhj3lq7FMt+p7cFHQ9xerrLBqyCF9j+4wu\naa6O9LN2ibvk27slU57i8h5Dmff1PI5WyLx/9ft1huh3xp83EXP7EDHXr/NnRK3oxInjzJ07lwED\nBrXiVTveXa5x4yZwxRVXXHCuZntSFAWDoeEeD71eh9N5NtZ77umcw+uETqFCluV/euKJi4oqPfG0\nAIyIHMW3+9dwmc/QeveHhfmRn1/WzlE17JbvbiIhOJHnR79MuiUdgEM56eTnl5FZnMnC/nfxxp5X\nOZVTcMGKi4qqkF6YTrgulvz8Mnr6SKw9vtr9Wree2MaQ8KHuxyNCx/CnnS+Rl1fqvnl5zHIMgI2H\nfib5dCrxvr3RVHpzfPFxDmYeZe5/r2HZ3m9q9VaeL7s0B6PDt9H3WFW1mHQmLFUWroyc2eLvyeLL\nF5OfX0YVHed7eoanf9Y644fFi9Al2rshAaNZO+dHFqy5mVHvjeGr2Svdhag8/fPWEiLm9tEVYm5p\ne6dt0VmXKLvd4U5SVFV1fwltq7H3+Mw6lILQDv4nSdIjkiQlSJIUd+arFa/fIRuU8bET2HxqU7s+\n5zfpy1nyc91iXo1xKA62Z29zDxU9XHgYg9ZARtkJqhxVFFYXcmXcFJyKk8OWuktvnOtUWSZVjip6\nBrp69CZ1u4I9eb9iqS6k3FbG9qyfGX/OuofjYydyuvwUx2uWyyizlVJkLSLcHEFKfjLpRTIJQa7i\nHCHmEBKD+zK5+1Q2nFzXYAz5lfnkVubQN6Rfo69do9EQ5h3O8MiRxPjFNnq8IDSiLdu7DtXW9Qrs\nw/dzfsCoNbJgzc1UO6o9HZIgdFoieTzHyJEj3cNTNRqN+ys3N7fTJi/Z2dlkZWU1fqAHNTb8ODg4\n5JIYNix0CgtwFfb6H7D5nK+LJsvyCVmWR7fGtVrb+NiJJOftoczWvHoZ2eVZvL77lRY954r0ZXx/\n/Ltmn7e/IIUKezmpBSmunkNLOiOjRnGy9IS70mrvwD70DOzFgcL9dc4/VZaJ1WkF4I87X2BI+FAi\nzK4lf/qHDiDcHMGPGRv4MWMDvgZfRkaNcp8b6ROFFJToTrRPlp4EYEaPWezJ283R4iP0Capd2fHK\nuKlszNyAU6l/pEdK/l78jQH08O/ZpNcvBSdyc+JFFwAWBGij9q6jtnX+pgA+mfn/KKjK58Ef7kER\na0YLQouIT+TnKCsrw+Fw4HA40Ol07qTm7bf/ye9+9zi+vp1vOEt2djYOh42h9Y9G6xBuv/32C+6f\nNWsWitKhbmIKlyhZluM9HYMnJIUOxNfox7asn5nWjKqEr/76V5Ye/IC7B9yPj6HpVY+dipPt2dso\ns5VS7ajGS+/V5HN3ZG+ju388J0tPcKz4KIcLD3NP0gNsPb2F9KLDmPVmAkyBXBaSVFM052xRGZvT\nxhVfjSHWL45re89hzbHv+OHGre62XqPRMKX7NNafWItGo2Fa/Ex02trzssbFTmDLqU3c0f8uMkpP\nEm6OYETkSB7b9Ah2xU7CectrTOg2kVJbKXvyfmV45Mg6r2dffjIDwgY2eQ7/51d93eT3ShAupCu2\ndyHeIXwxazlXrZjMkm1P8+61Hhm1Kwidmuh5PMdnn33KW2+9xauvvsLmzZspK3ONGzaZTLR89Kpn\nkx6DwYDNZqu1LTv7ND///LOHIqorNLTxhbwFoT1IkhQsSdJ7kiRtlCQpVJKkDyRJuuRXmdZpdYyN\nGc+WzI2NH1wjpyKbLw99CtDo8NDzHSzcT5mtFEVVOF5yrFnn7sjazvT4mXTzi2Pr6S3kVeQxsdsV\nNfu2EeUbjUajISl0AB/tf5/e/+nGs1ufAGBb1lZsThu9Anrz8o4lPH35kjprKU7pPp0fM39g/cl1\nzOw5q87zj4+dxNbTW3AqTjLKThDn152BYYOpdlYT5h1ea41IAD+jP5dHjeaHBoauJufvZWDY4Ca/\n/jMjYgThYnXV9q5HQE8+m7mMTw58yGvbX/N0OILQ6Yjk8Rz33Xc/ixcvZv78+Zw4cZznn19CcvJe\nrFYbaicd3hAQEEBJSYn7cXa2q4x9SkoKH3/8oafCuiilpa219J4g1PEe8CsQApThWo/2U49G1E7G\nxU7gp9NNH7H2dvI/SAobSK/A3shFzUset57+iQGhA4n2ieFI8eE6+x2Kg2+PrsJx3lq0qqqyM3sb\nl0ePISl0IN8cWY5Wo0UK7ku4OYId2T8T7RMDwJ1Jd/PB9E94fPiTLD34ISXWYtYc+5ZJcZP599QP\n+eHGrdw94P5634dqRxWKqtSa73jG6OgxlNlK2V+QQkbpSeL8u9MrsDdmvY97vuP5JnefxpeHPufx\nzY+yMr12z2FKXjIDw1qzSJsgNFmXbe8GRwzl31M/5PENj/PfIys9HY4gdCqdNnnU6XTY7bbGD2yB\n2NhuLFx4By+99BJbt27lyJH0TjvnMSIigtzcXOBs4qjV6rn++uvx8/PjiSce58UXn+eZZ55i06Yf\nPRKjoqiUlBQ3+fj33nuvDaMRurgesiz/C3DKsmyVZflpoEssKjosYjiHLGnYnI23q+W2MpYe+JDF\nQx6jb/BlpBUebNZzbcv6idEx4+gV1If0otrJo6Iq/GbTwyz6322sSF9Wa1960WEKqwsZGTWKgWGD\n2J71M/GB8Zh0JuL8urMvP5loX1fy6Gf0Z2r8DO4acB9h3uEsk79k7Yk1zOwxy90zqdXU/RPoY/Bh\nbMx4roybUu9wWn9TAIPCBvPT6S1klmXQ3b87Oq2O/qFJdXoxz5iTcCPTeszgaPER/vLLH9zb8yrz\nyKo4zcDwpvc8CkIr6rLtHcDU+Bn8c8Y/efCHe9ie1XFGYwlCR9dp5zx6e3tTUlLaqkMez6/66e1t\n5qGHHmbYsOF4ebVsIWWtVltrgfv2ZjKZag1b1WrPxjFmzDjGjBkHuJbLKCzMb/f4WkJUwBXakF2S\npIAzDyRJ6gN0jDVt2tiZojGFVQVE+Ua7t285tYmU/b/irwnhmt7XEWAK5GjxEaoclVwZN4WU/GR2\n5+5q8vM4FSfbs7axoN8dVDuqOFKcXmv/S9uXsO7E99yUOJ/Xd7/CnD43uucd7sjeRp/ABEK9QxkQ\nNhAVlT7BfQCI8+/Or7m/EH1O7ABajZZb+93OX3f9kXJ7OVO6T2s0xtcnvYVW0/AadGNjJrD19Gay\ny7OYHn8VAH8c91f8jQH1Hh9hjuAv4//O3tzdzFhxJVanFZPOREr+XgJMgcT792g0JkFoA122vTvj\n3mH3kp57nBu/vZYXx/yJhZctEsPCBaERnbbnUa/XY7VaW/Waq1d/R3l5Jamp+wBQVYUNG9aTkJCA\nl1fTCzqcS6834HDYWzPMZhs+fLi717EhWq2WkJCQRo/zhOzsbE+HIHQdS4BNQJwkSauAn4FnPRpR\nOwnxDkWDhrzKXPe202WnuP37W9h4YiMvbH+Wrw9/BcCp8lNE+ERi0BlIDO6L3Iw5j3vzdlNuL2Nk\n1Cj6BCVw5Lyex6UHP+RvE97g5TF/Jq8yj++OrQJcN40+S/uYKfHTAUiqGeqZEOLq7Yv3jwcgqmbY\n6rlu7nsb5fZyRkePI9Cr8SldET6RhJnDGtw/NnY8O7K2c7L0BHH+3QEYEDaI+IALJ4F9ghJqzfN0\nFcsZJD6sCp7SZdu7c/1+xNO8ccXbvLj9Oe5Zdwel1pLGTxKELqzDJ482m43y8vI6X3a7nZKSklpf\n1dUXt27PsWPHePfdd8jIyABAq9VRVVXFO++8TXFx04dVnkur1eB0enbI6+DBg9m6dStHjx6/4HHn\n9kq2t3/8400AiouL3YWKzvj8889rPRYftIS2IsvyWmAqrhL27wNJsiw3fz2JTkiv1RPiHUJ+VZ57\n25Nbf8fQiOFsXriZid2uILPM1TZmlZ8ipmZ4aGJwP7IqTlNibbyNPGRJ487/3cacPjfibwqgd2AC\n6cXp7tEEiqpQbisj1jcWf1MAdyXdw6u//oUqRxUbM3/gYOEB7h/4EADh5nCifWLcyWNcTfJ4fs8j\nuHr+Hhj4CAsvW9TyN+gcIyIvx67YqHRU0s2v6cvi+Rr9iPKJJr1IBmBf3l4x31HwmK7c3p3v+j5z\n2TB3M0eK05m8bDwp+cmeDkkQOqwOP2x1z57d/PLLznr37dpVe6jU1KnTmDZteoufy8/Pj9GjR7Fp\n0yYSEiT69Elg9uyr2b8/tcWJqcPh7BBrFAYGBrJ27RpMpuvo1s01paGwsJDU1H0YDAZGjRrVyBXa\njt1uJycnF1VVmTZtCrGxsSxffnYCu3Le+middf6p0HFJkrTknIcqcOYOxSBJkpBl+UUPhNXuwrzD\nyat0JY//O/E9mzJ+YNNN29FoNHTzi3Mnj6fLTxPt61qkvkdAT4xaI4cshxgZdXmt6721901m9bqa\n7v7x5Ffmc83K6czoMYtXJr4BuHriKuzl5FRkE+UbTaW9AhUVX6NrWaT7Bj7Ef49+w51rb6XEWsKt\n/W4nwifSff1PZn7J8F4DqS7F3QN4Jq7zPTPq+VZ7n8wGM8MiR7AzezsxDTxfQ/oESRwuklFVlV05\nO7lJrNkotDPR3tWvV2Afvp/zA8/9/CQzl0/mhTF/4M7+94gb1oJwHs9nNY0YMmQo/fsn1dm+efNG\nbrnl1lq/1CaT6aKeS1GchIWFcffd9/LJJ0u56aab6dWrNw6Hk5YuuWGz2S46rouh1FQrnDJlGkOH\nDufrr79i9epvUVWVoKAgkpKSiI/vSVRU3aFe7UFVVd544zUWLXLNM9i5s+7cqaCgYAoKCtzzW4OC\nLvlK4kL70+D6JR8BxALLcM39uQ64cJf9JSTUHE5+TfL4Y8Z6rup5NT0DegHQzS+Obad/Alw9j2cK\n0xh0BnoHJXDIcrBu8pjsShIfHPwIO7O3o9PqeW3SP93tdpRPNGa9D+nFh4nyjabcXg6AX03yGOgV\nxPKrv2X2N9PJKc/iP9M+rnX9pLCB+Jn8qKaMhKBEwrzDiWtGT+DFGBsznlNlmRh0hmadlxCUQHqR\nzNHiIxRWFzLivPdMENqBaO8a4KX34q8TXmNMzDge3fgw5bZyFg99zNNhCUKH0uGTR6PRiNForLXN\nZrPh5eVFYGBgqz6XTqfHbreTkJDArbfexqeffsKpU6e47bYFhIWFt+iaHeGO1ZnEMDg4mHvuuQ8A\ng0GH3e75efGff/4p48dPYPDgIQ0eM3r0KH75ZSczZ7oKU9x1113tFZ7QRciy/DyAJEnbgFGyLFfW\nPH4N15ygLiHcO9w9bPVUWSZJYQPd++LO63kcEjHMvc817zGt1rVUVaXEWkxqgWv41/6CfQwIG1ir\nTdRoNPQJSiC96DDjYydSbnMlj74GX/cxUb7RrLpmDXJRmjthrTd2czgH7jjS0pfebAsvu4sRkc1P\n/PoESXx68GN2Zm+nd2AfQr3FOrdC+xLtXeOu6X093npv7lh7KxO7XSEqIgvCOTp88lgfWU4jIKD+\nqnYX47bbFnD4sKvwQ0KCxIsvvnzR1/T0+pCKomC1Wt29n4qiUFZWhsGgw2z2beTstpWRkcHHHy9l\nzpw5VFZWMmHCRIKDg+scl59f4F5uRBDa2Pmf5I1Al+nqDjOHk1ORBbiSx5k9Zrv3dfPvTmF1IRX2\nCrLKT9cartk3uB+rjq7EqTjdlVGrHFXYFTup+SkApBakkBQ6kPP1CuzFsWJX0ldud813Nht8ah0T\n4xdLjF/zhoe2tTBzGBPMddeBbEyfoASOFB9me/bPjIzy3HQBQaCLt3eNmRo/g1v6LuCBDXezfu4W\nzAazp0MShA6hUyaP/fsP4PvvV+N0OtHpGi6n3lw6nY7jx49TWlqCv3/rJKeKoqCqqsd6IMvLy3nl\nlVfcybZWq8XX15e+ffsyaFDdQg379+9n69atTJw4kUmTrmzT2OLi4li3bj1lZWUcOXKY0tJidLq6\nNZzi4uIYO3ZcrW3nrgv57rvvUlZWhqI48fcPoE+f3vTpk8CAAXU/qApCI94DfpUkaTWgA2YBr3s2\npPYTZg4nNd9VbfpU+alaCVusn2uu9MnSE+RUZNfqBbwpcT7vpb7Ln3a+5J5beKaAzpHidMrt5aQW\npDBPuqXOc4Z4hVJkLQKgzFaGj8G33vUXLxV9giSqHFV8f3w1fxj7F0+HI3RtXbq9a4rnR7/M5GXj\neHH7s/x5/KueDkcQOoROmTxqNBpCQkKwWApbPJy0IQkJiWRkZNK/f+skjyaTCafTgV7fvHkxrSUw\nMJgnnniqzvaGhq0OGDCIvn0v4/XX/8748RNbNTlviJ+fH4MHDwVAUerOLe3ePb7W4/OPueeee93/\nLy4u5uTJE8iyTGxsbJ34Dx48yI8//ljnOfr2TeTKKycDEBDQusOhhc5DluW/SZK0EZiAa07QXFmW\n93k4rHYT7h1OXmUuJdZiymyldPM7u164r8GXEK8Qfs35BafqrNXzGOETyX+mLWXOqlkMjRzOjB5X\nUVyTPOq0OrZkbiKnIpv+oXXnr/uZ/MkoOwlAub3cPd/xUhXuHU6AKZASa7HoeRQ8qqu3d03hY/Dh\nncn/4aoVU5jcfSqTm7BOrCBc6jpl8gjU9Oa1/t3psrISwsMbXt+ruVwV6D0/77E5DAYDiYmJHD16\nlISEBE+H0yyBgYEEBp7tUT0/0YyKimbmzJl1zvPz80dRVLRaDSUlxSKB7KIkSTIAkUA+rl/cQZIk\nDZRlealnI2sfYWbXnMfMskygbuXSbn5x7Mjehl6rJ8y79o27y6NGcWu/21l++Ctm9LiKElsJJp0J\nKbgvXxz6BF+DH/EBPes8Z4AxkJKaddXKbWW15jteijQaDX0CE8goO0m8/4XXhRSEttTV27umGhQ+\nhN8Oe4LFPz7I5pt2iHnKQpfXaZPHioqKNqm6WV5e3qrXVZSOsVRHcx05coQ77midNdHam6qqbNiw\nnmnT6t4hzM3NYc+ePXW2S1ICYWGuPwgicezSPgfigDRql1juEh+mwszhFFuLOV5ylFDvMLz13rX2\nx/rFsSNrG1E+0e65jefvTy86DLiGrQaYAkkKHcCXhz5jWOSIeoejBpgCziaP9vJLPnkE6BtyGbF+\nsR2ioJrQpXXp9q45Hh7yKBsy1vHYpkf4aPpn4ndX6NI6X1ZTQ1XVNhlSabfbMZm8Wv26nU337t3b\nZchqW6iqqmLTpo3k5uZgMJyt1Dtv3jwSExNJTEysdfyXX37J4cPpHD6cjkaj4ZZb5uPnd2kPnRMa\nlAT0lWW5ZWvznEeSJC3wNjAAsAJ3ybJ8tDWu3RbCzREAJOftrTVk9YxufnF8d2xVg8MtQ7xCKKwq\nBKC4uogAYwBJYQP5LG0pSaED6j3Hz+hPma0UcM159L3Eh60CPDfqBVS1VX7EBOFitFp719nauubS\na/W8deW/mfTVGD5P+4T5/RZ4OiRB8JhOmzzqdDqqqqrw9vZu/OBmMJvNlJQUExIS0irXU1U8WjCn\npa699lqys097bP3Hi2E2m3n55T+Sk5NNeXkFJSWuXo3Tp7Pc3wez2ZvAQFcP87nFeHJysnnuuWd4\n7bU32j9woSNIA6KArFa63rWAUZbl0ZIkjQRerdnWIYV4haBBw57cX4mtZ73EOH/XtpgGlswI8Q6l\nsLoAONvzOKCmwmp9lVahpufR5vodrbB3jeQxwCRGNwgdQmu2d52qrWuJ+IAe/Gnc33hiy2MkhQ1g\nQMHD/x4AACAASURBVFjdooOC0BW0WfJYM5b+A6A7YAJextVQfQQowH7gwZbc8bLZbFit1lZPHAFy\ncnKIja17x72lFMXp8cQxNXUfRqORhITEJsWiKA6ATpk4arVnXp+GmJgY0tPTOXrUtQzAmX8Bevfu\nzfDhwwGIjT37Oqurq/j9759ot3iFDscHkCVJ2g9U12xTZVm+ooXXGwOsBZBleackScMaOd6j9Fo9\nId4hJOfv5bZ+C+vs71aTUJ4/F/KMYK9gLNWFrjUebSUEmgLpF9KfYK9ghkeOrPecAGMAZbZSnIqT\nclvXGLYqCB1Ea7Z3naqta6l50i2k5Cdzy+q5rLl+A3H+3T0dkiC0u7bseZwP5MuyfJskSUHAPmAv\n8JQsy1skSXoHuAb4prkX3r17F9ddN6d1o62hqipGo7HxA5tIW8+8oPa2bt06oqOjOX78ONOn1y0W\nc67OnDiece6cxWHDhjNs2PAmnzt06CX5905ouj/W/KtyttLVxQzp8gdKz3nslCRJK8uyZxeAvYAw\n7wjSLAcaGLbq+qB0oZ5Hh+Kg1FZCibUYf1MAZoOZtDuON3jjys/kD0CZrbTLzHkUhA6iNdu7TtfW\ntYRGo+GlMX/mVPkpbll9A99dt45AL7E0ptC1tGXyuAz4uub/WsAODJFleUvNtu+BqbQgedRoNGi1\nbbMOWGv3EnaE0apms5nhw4fjdNZdmuN8lZWV5ObmdtrkMSMjg6QkMSRMaBlZljdJkjQW6I9rlMSI\nc9qsligFzh2H2eCHqaAgM3q95282xQRGkWY5wGUxEmFhZ0MPC/PDy78fAP1iEmrtO8PgF+/6j9lK\ntaaCqMDweo87l8YntuZcBZumirjAmEbPOV9zj28vHTUuELEJrd7eNbmtg47T3kHLft6W3/wVkz6e\nxN0/LOB/t/4Pk97UBpE1rDP+joiY20d7xNxmyaMsyxUAkiT54UoknwFeOeeQcqDRxRRtNhs2m63W\nNkVRsFgsGAz/n737Do+qyv84/p6WXgkJTZqUQ5MqEBBp0sGCIgpiQ7Hr2nX9Kfay9ra6grr2Asq6\nIgKKNFERWZEichQUREAIJJBOMuX3x53EhJSZTCZzJ8n39Tx5yNy55TPD5Mw995x7Tvm5EyMjI4mK\nqt1gN8HvYmp+7TEtLY3jjz8eq9X3f3dMTBwffvgCAwcOqpcD5ixY8CFNm6bSokULs6OIekgpdT1G\nj4hWGOXWHKXUK1rrxwLc5VfAqcB8pVQ6sKmqFbOy8gM8RHAl2poAEO9OISMjBzC+jEp+v6bP9XSK\nPqH0cVkejxW71c4ve3Zx4MhBmqY0q3S9sopdxoXAX/ft4VBuFp3jI3xuU1bZbOEkXHOBZPN1/MYi\nyOWd32UdhE95V5vP279Hv8uEBacwfd4MXhj1cqWjSdcFs/9GAiGZQ6OmmQMt7+p0wBylVGtgAfBP\nrfW7SqlHyzwdDxz2tY8ffvietWvXVli+fv36CsvGjRtX6Rx+NeVw1KzSVP36nhrvL9jOPffcSpdX\nnsvG4MGDWbx4EZMnT67bYAGq7v284ooruOuuu3jttddCF6gGIiLq7RhVjcVFwEBgrdb6kFKqP7AO\nCLTy+B9gtFLqK+/ji2sfsW6VjLh6XCXdVgFmD7qvym0tFgtNolI4VHiIw0cPk+DHwDAOm4MYeyzZ\nR48Y9zxGSLdVIULkIoJX3tW7sq62UmNSeW/Sh0z4cBQPrb2POwfdY3YkIUKiLgfMaQZ8BlyltV7h\nXbxBKTVMa70KGA984Ws/HTt2omvX7qUtgm63m7Vrv2bMmHEkJZU/MYmMjKS42HfXzKoUFRXhdrtr\ntA+Hw1bl+rm5OTgcEbXKVFtOZzFbt/6I2+0hNjaOqKgokpOTSU5OrJArIyODQ4cOsmnTJi65ZJap\nuatS3ft95MgRtmzZRMeOHSkqcoY4mW8REfawzOVLMO8BrgdcWuujSqmSxwVAwP9p3gHBrgxGsFBJ\njUkj1hFHUmRg9/GkRKWQWXCII0cPk+TnqKIJkQlkF2WTV5xLnKPxtPwIYbKglXf1sawLhg5JnXhj\nwvtM+fhUWsa3YmaPWWZHEqLO1WUzyB0Y3VJnK6Vme5f9DXhWKRUBbOWveyKr9NNPWykoKCi3rGnT\nprRp0ybgLqbFxcWsWbO6dAqHshXTvn37Vlg/NzeXffv2kpOTTV5ePllZmeTm5nmfLbnPvOQe87/u\nObdYLFx+uX9lqcfjweVyYbPZqnxd27f/zNatW4mJiSE6Orq0W2lKSgodOnSosP769evZsGEDXbt2\nJSIigoMHMygsLKRLly7ExkZXmGfs0KEM/vzzT9LT02ndOngjzoIxCupHH33E9u1/jXhqsVjweDxM\nnTqVNm0qTgswf/48fv99N05nMW3btmPixEns3fsHnTurCu9RZmYmH3wwn44dO3HNNdcFNbtoVFYp\npZ4A4pRSZwCXActNzhRSXZp0oW9av4DL15KWxyNHj/g9JUViRCJHjh5uNPM8ChEmGn15FwwDW6Tz\nr9GvctlnFxFhjWBGtwvNjiREnarLex7/hlFZPNbwmuxn+vQZQclT1iuvzKVfv36cf/6FPk+QioqK\neOCB++jYsRMdO3akZcuWjBw5klatjsNisVTbElYTd955B6tXr8bpdPLoo49y0kknlXve5XJhtdo4\n4YSe7N+/n0OHDnH06FEA0tMH06xZswr7PPXUVpx66umVHq+y3HUxSE7J1BnZ2dlMmDCB1NQ0v7e9\n9NLLSn+fP38evXqdwPnnn8+TTz5dYcCkxMQkbrnltuCEFo3ZLcAs4AfgAuBT4F+mJgqxUW3HMqrt\n2IC3T4luyqGCgxw+epjESJ+3tQOQEJlodFstzpHRVoUInUZf3gXLxONP5YVRc7lq2SxsFhvTugb/\n3FWIcBH2N2C9885bFVoek5OTOeOMMwMecfWSSy7lyy9X889/PofH4yltAStpeTz55GGl60ZERPB/\n/3cXe/fuIScnm3379rF164/k5+fzV4tj2X8p/d1qtTBr1uU0bdrUZ6b7738Qp9OJ3W6v9HW53R5+\n/XUHubm5xMbGEh8fR3JyEg6HA7fbyb59e8qtX1BQwOuvv47NZqNLly44HA7y8vJwOBwMHTq09DWX\nyMnJ4dtvv8XhcBAXF8eoUWNqPfhQSW4wRnF95513yMvLKz2uxWJBKcWkSZMqbLdt2zYWLVoEgMvl\n5LjjWrNhw0ZiY6PJycmusP68ee+TlZVFx46d6NHjBDp37lzr7I3V5ZdfxjvvvM2QISdz8skn89BD\nD5odKWS01i6l1LcY8585gWVa6/rX19hETaKasD//T/KdeX63PCZEJHCk6IgxVYfc8yhESEh5F1yn\ndzwTt8fN1V9chsVi4dwu55kdSYg6UWXlUSm1QWvdRylV2VDLHq11SEaB6dmzN3FxceW6li5Z8ilF\nRUUBV24cjghGjhxVYXlubi4rVnxRrvIIEB8fj1JdqthX1S2Pe/fuYf78eVx55VU+M1mt1mrvLbPb\nHYwaNcbnfkrExsZz1VXX4HQWc/fds4mMjOScc6aRmJiI1WqvkDs6OpYOHTpRUJDPyy/PZfDgIUGp\nPJZIS2vGZZddXulzJRXMsjp3VnTurCosr+r9njJlKkeOHOHuu+9i+/ZfuPXW22sfupF66aU5vPTS\nHLNjmEIpdTNwOfAxYAM+UUo9pLV+1dxk9UdKdFM2HzQGWvT3nsfEyEQy8g/gdDvlnkchQkTKu+Cb\n3GkKLo+L65Zfic1i42xV+YCFQtRnVVYetdZ9vP9WaAZTSoVsQpuffvqRAQPSSx9brVYcDkdQKzYl\nfvttR1CnePjjjz9o1apl0PYXCJvNTvPmzRk9ejQJCXE0b9683PN33fV/LFmyhPHjxzNz5ky++eZr\nrr/+epo0aWJS4sC5XC46dOggFUdRG5cD/bTW2QBKqXuBrwE5mfJTSlQKvx3ZARj3MvojISKRPbl/\nABAv9zwKESpS3tWBKZ3Pwe1xc/2Kq7FarJzVearZkYQIKp/dVpVS32itB5V5bAPWAyfUZbASbnfF\nhs/gz8VoiI2Nw+kMXo8Nt9tNhw4dg7a/QBj3ZToYNmxEha6tAHfffS/33fcAFosFt9vJnXfOrmQv\n9UNcXBwXXniR2TFE/XaI8qMN5mJMfi381CQ6hczCTKwWq9+D3yREJPJd7joAuedRiNCR8q6OTFXT\ncHvcXLf8SpKjmjCyTcXebkLUV9V1W12Od3CbY7quuoD/1m2sv9RFC2NV7HZ7UCuPdrudoqLioO2v\nLtjt5T8C+/btqZOBc0IhIiKCqKiQNYqLhulX4Gul1LsYZd1kIEcpdTdGd/2qJzkUAKREGfd4J0Qk\n+D1pdkJkInu9LY8y2qoQISPlXR06t8t57M3dwzVfXM6Kc76mWUzFgQ2FqI+qa3l8HXgDeAYwbe6D\n7Ozs0ikswBj9NJgVvLKysjIrTF9RGy5XeNx3XnYaD7fbyfr16yksLCQnJ6fcADajRo0iOTmwud2E\naCB+9v6UXLX6nIrz8IhqNIlOAfB7sBxj3USyjmYRZYvCbg37cdyEaCikvKtjf+t7E1/+sYqrl13G\nvFP/4/cFNSHCWXXf0sO9/24EHsYYwtkJjAN+xKhc1rkmTZrwxRefl45A6na7mTYtOCNYeTweDhw4\nwL59eyguLmbHjh3ccMNNAe2rqKiIzMxM9u79gz///JPMzEzAwtlnnxOUrP4oGTm2hNttVF5nzLgA\n+Gsqjm7dCikszCc2No7Y2DisVisej4ekpKQKLZHh6ODBg7z//rulj886awotWxr3liYm+n/CKsSx\ntNb3KKXigA7AZiBGa51rcqx6pam35TEp0v8LUQkRCYC0OgoRSlLe1T2b1caLo19mxPuDeX7D01zX\n90azIwlRa9UNmHMxgFJqBdBTa33Q+ziZEHZbnTjx1NJpNMC4hy/QKTqOtXnzRpYvX47VasVutzN0\n6NCAKk/ffvsNTz/9NMnJyaSkpNC5s2LmzJkcd1zroOT01yuvzKWoqAir1YrFYsFut3PjjTdXWK9D\nhw5Bm5/SDJs3b2TPnj106NCBCy+8qF5UeEX9oJQ6BXgJo2wcBGxSSs3QWi81N1n9kRxlDLaV4Occ\nj0DpfJByv6MQoSPlXWg0j23BsyNf5MIl0xnU8iT6Nx9odiQhasWfs+6WQFaZx3lA8yrWDbolSz5l\n586dREYa97Ll5+dz4YUXExdX+5OMnj1707Nnb8Bo0Zw37z3Gjh1f48rpwIGDePfd0jGFyMw8xKJF\ni9ixYwePPPKPoFV2fbn00svKPd63bw8vvPB8pRXI+mzEiFMYMeKUcsvcbg9Wq4UjRw5L66OojYeB\nk4FPtdb7lFLDgXcBOZnyU5Q9ilhHnN/TdIAxYA5Iy6MQISblXYiMbjeOS3pcxhWfX8LyqWtq1K1f\niHDjT61mEfC5UupqpdR1wBfA+3Ub6y9//vknAwakl/6kpKTw+++7OHLkSIWfwsLCgI9jtVqJjY3F\n7a59a1yTJimMGTOW5OTkSkeLDZUWLVoxbdo0044fam63h02bNlFUVGR2FFF/WbXW+0oeaK1/5K97\ngISfUqKbBlR5lGk6hAgpKe9C6M5B95Ic1YSbVv4tqONrCBFqPlsetdY3KqWmAMMwCpXHtNYf13ky\nr4MHD/L222+WW/bdd99Vuu6YMWMZO3ZcwMfyeDxBbSX0eNwha3UUhtWrV9G//wCzY4j6a7dS6lQA\npVQScDXwu7mR6p+UqCalFUJ/SLdVIUwh5V0IRdoimTP6VU6ZP5Q3t77GBd0vNjuSEAHx62YxrfUH\nwAd1nKVSKSkpjBw5qrTb6pYtm+jR44RK7ycsWSdQVqsVl8uN1Wqr1X5K2Gw2XC6XVCBDKCoqmry8\nvJBO8SIalCswRphujTGM/XLgsmq3EBUMaJ5Oj6b+TwWcIJVHIcwg5V2IHZ/UkceGPcUNK66hTUJb\nhrceaXYkIWos7EcaOe6449i4cQMjRxoTrObl5XH88R2Ijo4O+rGKioqCOvjK0aPB3Z+oXnZ2Nrt2\n7ZLpRkTAtNb7gXPNzlHf3T/kkRqtH2WLwmF1yD2PQoSQlHfmmNL5HPbm7uHCxdN4f9J/SG852OxI\nQtRI2NdsRo8ey0cfLWDp0sXYbDaio6PrpOIIxkiuZae6qI2VK1fQuvVxQduf8O2VV17m7rvvkZZe\nUWNKqd/KPPRQfo4zj9b6+BBHalQsFguJkYnEOaTyKERdk/LOfNf1vZG84lymLzqbBacvpHdaX7Mj\nCeG3sK88ApxxxpkhOU4wK3q7d//O7Nn3BG1/gfjjjz9wOotK53ds6NxuNykpKWbHEPXTCO+/szG6\nb70GuIDpQK1PpJRSk4EpWuvgTFLbACVEJBIXId1WhQgBKe/CwO0D7iK3KJdzFk7mozMW0zWlm9mR\nhPBLvag81kfBum+yNvbu/YOmTVPNjhEyl19+udkRRD2ltd4JoJTqpbWeWeapJ5RS39dm30qpZ4Ax\nwIba7KehS45KJrEGg+wIIQIj5V14sFgs3D/kEfKd+Uz5+DQWTl7C8UkdzY4lhE/Sv6+OhENv1ejo\nWPbu3WN2jJBJSEgwO4Ko/yxKqdIRDJRSE4DiWu7zK+BKyncNE8d4esQLnNNlutkxhGhMpLwzmdVi\n5fFhzzCk1cmc9fFp7M6RwW5F+JOWxzIa2v2JVqvV1HkmhaiHLgHeUEq1wDj52QXM8GdDpdQlwPXH\nLL5Iaz3PO/m2qIZq0sXsCEI0NlLehQGb1cbzp8zhgsXnctWyWfz3jMVYLdK2I8KXVB69XC5X0CqP\nu3btJBwuuh048KfcAyhEDWitNwAnKKWaYgwccagG274CvBLosZOTY7Dbze/uXpXU1PAdzCZcs4Vr\nLpBsQsq7EuHyeXtzyut0+WcXFu35kJl9Zla7brhkrgnJHBqhyCyVR6+srEwSE2t/v01xcTGLFy/m\ngQceDEKq2tm2bRt33XW32TFCwmo1v7IuGg6t9cFQHzMrKz/Uh/Rbamo8GRk5ZseoVLhmC9dcINl8\nHb+xaczlndmft7LsxHHHwLu5eenNDE4ZSUp05Rf/wymzvyRzaNQ0c6DlnbSLexUWFgZlYvm9e/fQ\nrVs3HA5HEFLVzrBhw7DZwuPKXl0rLCwkMTHJ7BhCVMXj/RFCiIZOyrsAXdhtJu0Tj+feb+40O4oQ\nVarzlkel1EDgEa31CKVUH2Ah8Iv36Re11vOq276wsJCjR4/6dazIyMiAK4A2m42CgoKAtj1Ws2bN\ngrKf2urSpfHcQ/TBBx+glGLUqNFmRxGiAq31KmCV2TmEEKKuSXkXOJvVxmPDnmbMB8M5V53H4FZD\nzI4kRAV1WnlUSt2KcfN1rndRP+BJrfWT/u5j1aqVfPbZUr/WHTNmLGPHjqtxToC0tGasW/dtQNuW\n1bZtOwoLC/jpp6107Wr+nD379u1pFPM8Tps2naeeepKUlCb06dPP7DiinlFKrajmaY/WemQ1zwsh\nRL0h5V14OyG1F5f2vIJbV9/A8qlfEWGLMDuSEOXUdcvjduBM4E3v435AZ6XU6Ritj9drrXOr2hhg\n2LDhpKcP8utgkZGRAQcNZvfOqKhoCguD04pZG1arHbfbaXaMkLBYLIwZM5aff/5FKo8iEPdW85x0\nvxJCNCRS3oW52/rfwcLtH/HCD89yfb+bzY4jRDl1WnnUWi9QSrUrs+hbYI7WeoNS6g7gbuCW6vYR\nFRUVlHsR/eHxBFZmFhUdZe/evYCFyMhIoqOjsFrldtJQy8/PJzo62uwYoh7SWq8s+V0p1ReIxRgy\n2Q60Q7pgCSEaCCnvwl9cRDwPnvwoV35+Cad3PJP2icebHUmIUqEebfU/Wusj3t8/Ap71tYHdbg3Z\n/IsWiwWHo2YtkHl5eaxevZqvv/6a7Oxs3G43LpeL7t2707FjB5KSzB3EpajIVeVrqulrDRdV5bbb\nrRQVQURE+A4iHM7ZBCil3gAGASnAVqA3xqTXr5qZSwghgk3Ku/A2of0khrUewe2rb+K9SQsa3Fzk\nov4K9ZnsUqXUtVrr74BTgPW+NnA6QzvJfXGxq0brb926lQ0bfmD27HvKLf/oowU89dTT3HnnXUFM\nVzM//riZyMhImjdvWeE5h8NW49caDqrL3aZNWxISEigqCs+uuhER9rDNVp2IiEZ1v8VQoDPwHH9d\n3PqneXGEEKLOSHkXxiwWCw+d/BjD3x/MzKXn8/iwZ6qcvkOIUApV38qS/qBXAk95b9YeBDwQouP7\n5HQ6A+pq6vF4iI6OxuksJj8/r3Rk2KSkJNOn69i1a1ejGCynRFpaGs2aNTc7hqjf9mqti4CfgJ5a\n6x+BxjfxmxCiMZDyLsy1jm/D52ev4s+8vQx7P53lvy8zO5IQdd/yqLXeCQz2/r4BCNtxhwPpEtC6\ndWtiY2N56623aNu2LXa7nf379/Prrzs466wpdZDSf263u9HM8yhEkOxRSv0d+AJ4VCkFxv1AQgjR\n0Eh5Vw8cn9iBhZM/4+n/Pc6MT6eyZv8V3Nz7TmIcMWZHE42UjOpSCx6PhxYtWnDeeefRunVr9u/f\nz++//05sbCznnjstLOYc9HhC2+1XiHruEuA3rfU64EPgXIweE0II0dBIeVdP2K12bu5/OwsnL2Xp\njqWMnj+UTRk/mB1LNFJSefSyWCw1Hm3VYrGQn5/Ppk2bKSoqon///px33gzGj59Abm4uP/30Ux2l\n9Y/VasXlksqjEL4opUr6OycB3yil2gAfA9cBO0wLJoQQQSblXf3Vr1l/Nly+gfSWJzFpwRiW7lxs\ndiTRCMnQj7W0Zs0aNmzYQMuWrXjrrbfo06cPp512Ops2bSQvL5+uXbuamk9G5xLCL68AE4HVVJzn\nzAPIOOlCiIZCyrt6LDYilieGP4NKVly85DyeGPYs07rOMDuWaESk8lhLv//+O4MHDyE9PZ2MjAxe\nffUV2rdvj8PhCNn8lFXp168fzZs3ngFkPvnkE847TwpQUXNa64neX/tqrTPLPnfMXLVCCFGvSXnX\nMFzW6ypSopty3fIrySjI4No+10uDgQgJqTzWUpMmTfj11x1069aN1NRUxo+fwLp16/j5Z82gQYNN\nzZaamkpMTOO5oVprbXYEUU8ppVpjdONfpJSaUOYpB7AI6GJKMCGECDIp7xqOszpPJTmqCTOXnE9G\nwQHuHfwgVovckSbqVthXHgsLC0unv/AlMjIy5K19o0eP5oUXXmTLls0MHnwSPXv25OjRQpYt+5zC\nwsKQZhFCBOw+YDjQElhVZrkT+MSMQEIIUUekvGtARrYZxYLTFzJ90RQOFRzkmREv4LCZO1WcaNjC\nvvK4atVKPvtsqV/rjhkzlrFjx9VxovLi4+O58cabAGP0VbfbTf/+A/jHPx6lX79+Ic0ihAiM1vpi\nAKXUbVrrf5idRwgh6oqUdw1P32YnsnDyZ0xdeAb3rZ3N/Sc9bHYk0YCFfeVx2LDhpKcP8mvdyMjI\ngI9T05FWK9vOYrGUzqu4ffsvdOnSlbg4mW83lJxOJ3Z72H+sRfi6GJCTKSFEYyDlXQPSKbkzL4ya\ny5n/ncRUNY0TmvY0O5JooML+LDsqKiokXVFdLldpxa8mLBYL+/btIz8/D7vdTlRUNNHRUURFRRMZ\nGVEHSUVVxowZw7///QqzZl1udhRRf/2olJoNfAsUABbAo7VeXdMdKaUSgbeAeCACuFFrvTaYYYUQ\nohakvGtgBrU8iSmdz+G2VTfyyZmfyf2Pok6EfeUxVAJtsdqyZQvLl69g586duN0uioqKiYmJ4aab\nbiIpKbkOkvrviy++YMqUs4mKijY1R6h0796DtLQ0s2OI+i0FGOH9KevYx/64Afhca/2sUqoz8C4g\nfdmFEOFCyrsGaPag+znp3X68/dMbnN/tIrPjiAZIKo9eTqczoJbHN998k8suu5w2bdqWLsvIOMDz\nzz/PE088QdOmqcGMWSO//fYbDkfjav1s1qyZ2RFEPaa1Hh7E3T0FlIz25cC4si+EEGFByruGKTUm\nlf9Lv4f7v5nN+PaTaBrd1OxIooGRyqOXx+MJaH6c6OhoiouLKSgowGo1ugfEx8cTFRVFRETg92AG\ng8fjCahCLERjpZQ6GbgFiMUYyt4GtNFat/Ox3SXA9ccsvkhr/T+lVHPgTeBvwU8shBCBkfKu4ZrR\n9ULe/elN7v9mNs+MfMHsOKKBkcqjl9Vqxe1213i7IUOG8Omnn9K+fXvi4uJwu91s3ryZ448/npgY\nc7uL2mw2iouLcThkyGYh/PQyxgASFwLPAhOA731tpLV+BXjl2OVKqRMwum/dpLX+srp9JCfHYLeH\n78We1NTwHfwrXLOFay6QbAKQ8g6on583fzLPPWMOA+YO4KpBlzOkzZAQpKpeQ32fw00oMkvl0Ssm\nJoaCgpr3shg1ahSdO3dhx47tHD58GKvVytChQ+natSt2u7mVthYtWrB7926OP/54U3MIUY8UaK1f\nVUq1A7KAWcD/AtmRUqobMB84W2u92df6WVn5gRwmJFJT48nIyDE7RqXCNVu45gLJ5uv4jUijL+/M\n/rwFwt/Mre2duLjHpVz238tZdvaXps792JDf53BS08yBlndSefSy2WwBtTyCUUlr0aJFuWUff/wR\nXbt2C0a0gKWmpvLjj5ul8iiE/wqUUk0ADaQDK4CYAPf1EMaog88qpQAOa60nByWlEELUnpR3Ddzt\nA+5k8LsnMmfTi1zd5zqz44gGQiqPXi6Xq/SexWDIzc0N2r4C1a1bd4qLi8yOEVI//bSN3bt/5+yz\np5odRdRPTwLzgMnAemAGAV6J11qfEcRcQggRbFLeNXAJkYncd9JD3LjiWhIjE5ne9XyZvkPUmnyC\nvDweT1Arj8Z0SeaKi4snMTHR7BghZbVaA+p+LASA1no+MFprnYMxzPwM748QQjQoUt41DpM7TuHu\nwQ9w7zd3MeHDU/jhgM/bWoWolrQ8irD2yy+/8Pjjj7Jnz16mT5/OtGnnYrEYN9mXjJA7Z85LBwFF\nBAAAIABJREFU7Nq1i6ZNmzJixEisVgsulwubzcbGjT9w5MgRBgwYSFRUlMmvRoQrpVQr4DmgM7BG\nKXW71vowfgweIYQQ9YmUd42LxWLh4h6XcmqHM3hw7T2M//AUpnc9nzsG3k1KdIrZ8UQ91OhbHjMz\nM/ntt9/44YfvA5qqoyy3201mZibr16/j8OGsICWsGY/Hg9NZ7M3jNCVDMJTcf3rbbbdyzjnnMm/e\nfBYuXMj27dtLn7dYLMyfP49ff/2VAQMGsGfPHl5//TWsVhs2m42MjAwefPBB/vWvf5GfHx4354uw\n9W9gG8aw9VEYc5YJIURDJOVdI9Q0uilPjXieRWd+zuaMjQx+py/z9XtmxxL1UNi3PBYWFnL06FHf\nKwKRkZF+ty4tXbqYvLw8kpKSSExMpFu37rRt267abTweD/v3/8m6dd+SnZ3j7ebq8T4HFgskJSXR\nsmVL7r//Qb9y1MTu3bv4448/aN26Lc2bN8dut+Nyudi4cQPr16/H5XIB0KlTJ0aOHAlAixatgp6j\nNnJyctiyZRO//74bMK6InXvuuRXWc7s9PPXUE2RmHmLjxh/49dcduFxOnnvuOZ5++hnsduOjW1hY\nQKdOHRk3bhwLFiwgISGeSZNOBeDJJ5+gSZMmxMfHB7lLsmiAWmqt7wBQSi0DNpqcRwgh6oqUd41Y\n32Ynsvis5byx9d9cv+Jq8orzuKjHJWbHEvVI2FceV61ayWefLfVr3TFjxjJ27Dif6xUVFbFv3z5O\nPPFERo8eW1oR8WX+/Hl88cUyevbsSfv27bFYjMpjdHQ08fEJ9O7dh7S0NOLj42vdinkst9vNnDlz\nmDRpEt9/v579+/ezc+dOhgwZwsaNG7n++uuJiIgoV1n0eDxBzVBbTqeTiRPHc9VVVzNs2DCMt6jy\n98lmszF27Di2bdOcffZUkpOTycvLIyoqCpfLVfp/NmTIEK699lp++GEje/fu5c47XyQhIYHnnnuW\noUOHERcXy+rVq4mLi/OZLy8vDzCmbQn2/58Ie6UjS2mti5VS/l2xEkKI+kfKu0bOZrVxcY9LaRqd\nyhWfz8RisXBh95lmxxL1RJ1XHpVSA4FHtNYjlFIdgdcAN7AFuFprXW0NZ9iw4bhcTtavX4/FYqFL\nly6MGVN5BTEyMtKvTBEREZxzzjS2bNnESy+9iMfjwePxkJyczIwZF1S53dSp5zB16jl4PB6ysrLw\neDw4HDays3M5eDCDNWu+JCsrk9xcoxISFxfLDTfc5FcmX6xWa5WtmePHTyz9fd++PaW/r1mzhgMH\nDnD11dcGJUNt2e12Vq9eU2G5213+I/Daa//m888/Z8uWzXTp0oXk5CbExsbidhv/RxaLleJiJzab\njTlz5jJwYDpdu3ajefNmrF37NX/88Tv/+teLXHrpLH755Wd27tzJ6tUr6dfvxGrzzZ07h+JiZ2kX\n1+nTpxEbG8eePXtwu914PB5cLhf/+c8CEhLKD0RksVg477zz6NxZ1fJdEiaRqwVCiMZCyjsBwKkd\nTgdeNSqQWLig+8VmRxL1QJ1WHpVSt2KM3FUyb8WTwB1a69VKqReB04GPqttHVFQUY8eOp1u3Hlgs\nkJLS1K9WJF+io6Pp2/dE1q79moMHD+J2u/3u2mixWGjSpAkADoeNhIQkjjvuuNLnDx8+zOLFi4iM\nDP0ALVbrX/+lxx13HLt37w55htq66KKLmTZtOh6Ph1GjTsHpNO7dXL16FQ8//HC5luJ169bxwgsv\noFQX8vLyGDlyOAMGDOTaa68jKyuLI0eOkJeXR2ZmJtnZ2ZW2YsfGxjF+/Hhmzbqs3PK9e/fy/vvv\n0arVcaXHtNttnHba6fTr16/cui6Xi8TEpOC+ESKUuiulfivzuGWZxx6ttUyWKoRoKKS8E6XKViAB\nqUAKn+q65XE7cCbwpvdxX631au/vi4Ex+Kg8gtGFsV27dkEPN3/++/Ts2ZOpU88N2j1x27b9xNq1\n33DTTbeQkJAQlH0G6tNPP2X27HtMzRCoklbk22//OzNnXozb7WbgwHS6dOlCQUEB999/Hw888CC3\n3XYbzz//HElJSeTl5dGq1XG8+eZbpd1OP/roP3z33XdMmTKVoqIiRo4cVeFYDoej0opfYmISXbt2\nq9sXKsJFZ7MDCCFEiEh5J8oxKpCvcMXnl2CxWDi/20VmRxJhrE4rj1rrBUqpdmUWle0qkQuYNgnh\n0aNHiYuLY+zY8UHd7zfffM3s2ff4fR9lXbLb7TgcDrNj1MqECRNIT0+nqKiIpKQkHA4HLpeH886b\ngdVqZfToMURFRZGbm0t2djZdunQBjJZAi8VC69ZtsFptuN1uIiIiaN++vcmvSIQjrfVOszMIIUQo\nSHknKnNqhzPweDxcuexSMgsOcW3fG7BaZLBBUVGoazjuMr/HA4d9bWC3W+tk8JLs7ALi4+NxOGy1\n3lfZfURGRhId7d+9l3XNYrFU+/qC8drrno0WLZqVWxIVFUHv3j1LH59yijGy7KFDh/joo4+IjPyr\nwjxo0MDQxPRTRIT5FxWEEEIIIY51WsfJxDpiuWrZLNb9uZbnT3mJ5KgmZscSYSbUZ7IblFLDtNar\ngPHAF742cDrdvlYJiNNpDIBSXOyq1X4cDluFfdR2n8EyceLEKrNUlrs+qC630+nG7fZQVBSe81tG\nRNjDNlt1IiIizI4ghBBCiBA4pe0Ylk39kllLL2TU/KHMHfMafZtVP+CgaFxC1R5dMpzmTcC9Sqmv\nMSquH4To+BVERERQXFxs1uFDomXLlmZHCKn4+HjGjfM9VYsQQgghhKhc6/g2fDx5KWPbjee0/4zj\nlc0vhd30b8I8dd7y6O1bP9j7+y/A8Lo+pj8iIyM5ejT4UxvJ35Z5HA5Ho6swCyGEEEIEW4QtgodO\nfoz0FoO5fsU1fL33K+4/6WFaxrXyvbFo0BrtnbB2u710CohgkiszQgghhBCiITit42Q+P3slBwsy\nSH+7D/d+fRdZhZlmxxImarSVx4KCfKKjo4O+X6k8CiGEEEKIhqJDUic+Ov1TXh33Jst/X8aAt3vz\n7PdPUeAsMDuaMEGjrTx6PMGv6Hk8njrpCiuEEEIIIYRZLBYLo9qOZfnUNTw45B+8/uMrDHy7N6//\n+CpFriKz44kQarSVx9jYWHJzc4O6T4vFQlpaGp9+uiio+w3Uq6++anYEIRotpVSsUuq/SqlVSqnP\nlVJyQ64QokGS8q7xsFltTFXT+Hr6/7i693U8uu4hBr7dm1e3zKXQWWh2PBECjbbyuG7dWlq3bh20\n/Xk8Hj799BNcLhdjx5o34ufXX6/h+eef5aWXXqRp06am5TDD5s2bWLt2rdkxhChxKfCd1noY8BZw\nq8l5hBCirkh518hE2iK5vNfVrD9/M1f2uoan1j/GgLd7MXfTi9KdtYFrtJXH3bt3k5qaGrT9HTx4\nkLVr19KxY0dsNlvQ9lsTHo+HxYsXc/rppzNr1ixOO+00U3KYZdu2bcTGxpodQwgAtNbPAA95H7YF\nskyMI4QQdUbKu8Yr2h7NZb2u4rsZm/hb35v454Zn6f9WT5buXGx2NFFH6nyqjnDVtWtXFi5cSN++\nJxIfH1/r/aWmpvL3v/8fH3wwj4ULP2bu3FeCkLJmLBYL9933AC6XC6vVjttd/yakr43Jk8/kvvvu\nZciQk82OIhoZpdQlwPXHLL5Ia/0/pdRyoDswJvTJhBAiuKS8E5WJskdxyQmXMaPbhby08Z9cvOQ8\n/tb3Jm7p/3ezo4kgs4T76KBPPvl0UAPm5+ezePEimjZtyogRp9CiRQusVv8aYD0eD7/+uoONG38g\nP7/Au91f8RyOCJo3b87w4SNo06ZNMGP7JTs7m8WLFzFgwADatm0LQIsWlc/H43DYKC52hTJeUFSX\ne926b9m6dSs33HBjiFP5JyLCTlFR/avQx8XFWMzOUN8ppRSwSGvdsap1nE6Xx243p9eCEKKUlHe1\nJOWdAFj+23LO+eAcBrQawFuT3yI5OtnsSKKigMq7sG95LCws9HsE08jISKKioios37TpB5YtW8aN\nN97Mhx/O55prriMhIaFGOebPf5/58+eTmprKWWedxTXXXIfD4cCod5rf+7eoqIibb76R9PR02rfv\nQEJCQp1MRRLO5s17nxdffMnsGEIAoJS6HfhDa/0WkAdUe+UgKys/JLkCkZoaT0ZGjtkxKhWu2cI1\nF0g2X8cXNVdfyzuzP2+BqC+ZT4jrz+dnrWbm0hn0n9ufl0e/SfemPcyO5bf68j6XVdPMgZZ3Yd/y\nOHbsOM9nny31a90xY8b6HKxm69Yt7N+/nwsuuCigPC6Xiy1bNqP1NtxuNxaLBZfLhccDFm/9PSEh\nkVmzZhEVFfrK288/b2Pjxo20aNGCwYMHN6qWx/z8PGJjY0lKCs+rW9Ly2LgopdKA14EowAbcprX+\npqr1MzJywrYwDucv0XDNFq65QLL5OL6UdwGor+Wd2Z+3QNS3zIXOQu797u+8u/ld/j7wLo6Lb4Pd\nasdusWGz2rFb7bRLaM9x8cEbxDIY6tv7DAFVHhtmy+OwYcNJTx/k17qRkZE+1+nWrQc///xzwHls\nNhu9evWmV6/eQOWVmS+/XI3WP9OrV6+AjxOozp270LlzF4BGd89jTEwsFot874vwoLU+AIw3O4cQ\nQtQ1Ke9EVaLsUbx82st0S+zF8xuepshVhNPtxOVx4nQ7KXY7ySvOpVNSZ0a0OYURrU9hUMshxDhi\nzI4uqhD2lceoqKhKu6LWRl1XMOLi4ti+/RdTKo9CCCGEEEKEk/O7XcT53S6q9LmdR35j5e7lrNj9\nBbM+u5gi11EGtTyJO9PvoVdan9AGFT6Zf7NeA5STkyMVRyGEEEIIIXxol9iei3pcwuvj30HP3MkH\np31Mq7jjmLBgFM9+/yQud/27zaohk8pjHTh6tJDoaGluF0IIIYQQwl8Om4P0loN5euQ/eXXcW/xr\n4/Oc+fEkduf8bnY04RX23VbrQnFxMWBMvbFgwQccPnwYh8NBhw4dOOmk2s8RmJCQwNy5czjxxBOZ\nNOnUWu8vUBs3bmTDhg1MmDDJtAyhlpWVxdNPP0Xr1m1o1iyttItyVFQ0/fr1q7B+fn4+GzZ8X25Z\n7959admyZUjyCiGEEEKIisa2G8/Kc9Zy/fKrGPH+Sfxj6BOc1Xmq2bEavQbV8lhYWMjSpUsoLCys\ndr2WLVvy7rtv89RTT+B2uzl6tIguXbrx1VdfkZ/v//DRBQUFvPfeexQUFJRbPnDgIM4551x+/HEL\nn3/+GQB5eXnMmfMSeXl5NX9hAW7fu3dftm/fzhlnnMYjjzzEPffM5t5772b79p/ZvXs3+/btKffz\n2GP/4L777uHRRx9hx44d1e47NzeXJ598gtzc3IBej79qepzExCSuuOIqevfujc1mx2q1eX+suN2e\nCj8eD2XWsWGxWPngg3n885/PceTI4Qo/O3f+xurVq1izZk3pz1dffcXWrT9WmicnJ4eHHnqQnJz6\nNWKXaHxyc3N49NGHyM2t2Wc10O383Ucg+w/VNg1xv8ES7vlE/VOTz1Rl61a1fU2X1yRbXZePodje\nbGkxabw9cT53pM/mhhXXcMHiaaz+YyVuj9vsaI1Wg6o8Hj16lM8+W+pzXshevfqwdetW4uPj6dq1\nOytXriAmJoa0tDS++upLv49XUFDA+++/X6HyCMZAP7179+G3334FjBauuXPn1KhyWlYg21ssFoYO\nHcb333/Prl27aNGiBf369WP//gMUFBRgtdrL/cyYMYNRo0axf/9+3nrrjWr3nZuby9NPPxWSymNN\njxMREUH79sczcGB66U/J6LjHio6OLrdeevogzj//QqZPn0FeXh6PPfYojz32GE888QSPP/44Tz31\nFM8++yzZ2YdLfw4fziQzM5MjRw5Xmv/hhx+q8/dJiNrKzc3l8ccfqfFnNdDt/N1HIPsP1TYNcb/B\nEu75RP1Tk89UZetWtX1Nl9ckW12Xj6HYPhxYLBZm9pjFF1PXEGuPZcaiqQx4qxePf/cIf+TsNjte\no9Mou62uXLmcW265lZUrV7By5XLAqHhmZ2dz0UUzg3acL7/8ktmz7w7a/gJRWGhUbO+++x6Ki538\n8MP3/PbbrxQU5NO9e/cK63fr1p0+ffoRHR36OSrDgdX610i8LVq05L77HsBqbVDXWIQQQggh6p1O\nyZ15cfTLHDl6mAW/fMA7P73JY989zPDWIxnYYhBpMc1IjUkjLTqN1Jg0UqKbYrdUrOrYrDasFjm3\nC1SjrDw6nU4SEhI57bQzeOKJxwCjy2tycnAnl4+OjiYiIiKo+6ypmJiSgXsstG7dmtatjUlYK5uf\nUhgSE5NKf5d5I4UQQgghwkdiZBIX97iUi3tcypaDm3lfv8PafV9zIP8AGfkHOFiQgQdPtdsPbjmE\nk1sN5eTjhtM5Wcn5Xg00ysqjxWKhuLiYnJyc0g9LZGRk0Jv0CwsL8Xg8YfeB9Hg83kGD5KpLZY4c\nOcy+ffvo0qWr2VGEEEIIIUQVejQ9gR5NHy63zOl2cqjwEJkFhyq9N3JP7m6+/GMVb259nTvW3Eqz\nmOYMaTWU8e0nMrrdOKLtjbP3nb8aZeWxdevWfPLJx2zevJmJE09l7dq1REVFBXw/YlVatWrFrl27\naNeuXVD3G4gDB/7E7Xby/fff88477+BwOCgsLMTlcmG1WunRowczZ87EZrMB0KJFK5MTm8MYSMfN\nW2+9yU8//cT119/AyScPNTuWEEIIIYTwg91qp1lMM5rFNKv0+e5NezCm3XgAMvIz+GrPalb9sYJb\nVl1P8Qon49tPZGb/C+kZNwCHzRHK6PVCo6s8FhcXsW3bNq666hqioqJYtswYDXXp0sWkpaX5tQ+P\nx0NOTjbZ2dkAZGQcIDMzk4MHMzh8+DCHD2fhcrmxWAjZlA9OZzH79x/gzz/3cuTIEVwuF263m969\njYFi0tKak5aWhsWygTPPPJPi4mLsdjt2u/ERSE9PB+p/pbHsPYsAP/64hSVLllZYr3v37owbN67C\n8i1btrBkyRLatz+eKVPOpnNnVWdZhRBCCCGEeVJjUjmj01mc0eks/jH0SVbtXs6CXz5gyrwpRNmi\nGNV2LLGO2NL1y/YmtBD8noUWLMRFxJEYmUxSZBKJkUkkRSYR64j1ebzk4liysqqelcFhi6Brk261\n7hFp8Xiq7hMshBBCCCGEEEKA3PQmhBBCCCGEEMIPUnkUQgghhBBCCOGTVB6FEEIIIYQQQvgklUch\nhBBCCCGEED5J5VEIIYQQQgghhE9SeRRCCCGEEEII4ZNUHoUQQgghhBBC+GQ3O0C4UkpZgReAnsBR\n4FKt9Y4yz98AXAJkeBddBmwHXqxqm1CpLrtSqhnwXpnVewO3aa3nKKW+B454l/+qtb4khLHx5hsI\nPKK1HnHM8lOBuwAn8KrW+mVf/0ehVk32acDfMLJvBq7SWnvC4f325qsqd9h+xkXw+VHm9QeeACzA\nn8AMrfXREGes0d9YGOQ6D7gRcGGUW/8KYSYH8CrQFogEHtBaL6xkvTnAIa3130OYzQbMBToDHuAK\nrfWPZZ6vUN6HKptoeJRSfwdOBSIwyrjvgX8BxcDPGGWdRyk1C+N7zonx97LIpLwXAhd5H0YDvYB0\n4HlvtqPABVrrA2GeuQ3wMpAE2DAy/xpGmR3A6xhlpAuY5f33NcANbAGuDrPPRoXMWmvtfW46cI3W\nerD3cZ1llpbHqp0BRHj/E27HOGkqqy9wvtZ6hPfnF2Cyj21CpcrsWuv9JZmBO4D/AXOVUlHe50te\njxkVx1sxTigij1nuAJ4ERgPDgMuUUmkYrzMyDN7v6rJHA/cDw7XWQ4BEYFI4vN/efJXm9grnz7gI\nvirLDaWUBZgDXKS1PhlYgvHlFTI1/RszO5fXY8ApwEnATUqpxFDlAs4DMrTWQ4FxGCee5SilLgd6\nYFTgQmkS4Pb+f90JPFgmU1XlvRA1ppQaDgzylmvDgNbAbOAeb1kWCUxUSjUHrgUGA2OBh5VSEWZk\n1lq/XuY8bb031zMYFZkRwALgNm9jQDhnfhR4U2s9DOPvvEs4vc/ABMCmtT4JuA94CON77w5vuWkB\nTg/zzA8CKKX6ADNLVqrrzFJ5rNpJGCdIaK2/BU485vl+wB1KqS+VUrf7uU2o+MzhPRl8FrjSe4W+\nFxCjlFqqlPrCeyU91LYDZ2L8wZbVFdiutT6itS4G1gBDMV7nYjD9/YaqsxdifHEVeh/bgQLC4/2G\nqnNDeH/GRfBV93/bGTgE3KiUWgk00Vr/HOJ8Nf0bMzsXwCaMq+7R3udDWUmbj3GSDMZ3vbPsk0qp\nwcAA4CUqz15ntNb/BS73PmwHZJV5uqryXohAjAE2K6U+AhYCnwAbgBTveVA8UITxt/CV1rpYa52N\n8Xfd06TMACilTgS6e1vez9Vab/I+5cAo48I980lAa6XU5xgXs1YSXpk1YPd+DhIxPgf9tNarvc8v\nBkYB/QnjzEqpFIxK5PX8VZbX6fsslceqJQDZZR67vN26SryL8eU3EhiilJroxzah4k+OU4Et3tYk\ngDzgMa31WOAK4O1QZ9daL+CYExyvBP7q3gmQg/FHEy7vd5XZtdYerXUGgFLqWiBWa72MMHi/vfmq\nes8hvD/jIviq+79tinEF8zmML9NTlFIjCKEA/sZMzeX1I0bvji3AQu+XeKhy5Wmtc5VS8RgVyf8r\neU4p1QKjYnkNIa44lsnnUkq9jnER850yT1VV3gsRiFSMC6FT8H7XAr9gfO62AmnAKsLzc3cHcA+A\n1vpPKL3oczXwFGGeGePCUKbWejTwO3AbRmU9XDLnYWTchnER7VnKl4dlzzXDNfPzwCsYt0fkllmv\nTjPLSV/VsjE+5CWsWmt3mcfPaK0zvVdGFwF9/NgmVPzJcR5GN7QSP2MUqngrlIeAFnUZsgaOUP71\nxAOHCZ/3u1pKKatS6nGM7mtneReH8/tdIpw/4yL4qvu/PYTRGqS11k6MFsqwaXWu4m/MVEqpnhhd\njNpifNk3U0pNCXGG1sBy4A2tddl73adgXBD4FOOEbrpS6oJQZgPQWl+I0ao919v9GCov77OO3VYI\nPx0EPtNaO729JY5ifPcO0Vp3Bd7E6KoYVp87pVQS0FlrvarMsnMwxhyYoLU+RMUyO9wyHwI+9v6+\nEOM7I5wy3wAs0VorjPE/3sBo1S2RQOXnmuGUeQ3QHeNz8S7QTSn1JHX8eZbKY9W+wvjiRymVjtH9\nCO/jRGCLUirW23Q8EqOPd5XbhJg/OU7UWn9T5vFMvPc4KaVaYvzR7KvjnP7aBnRSSiV7+2wPBb4m\nfN5vX17CuK9icpmudeH8fteHz7gIvur+b38F4pRSHbyPT8ZoTQsXlf2Nme0IRteyo95K+AGMLqwh\n4b0f6jPgVq31a2Wf01o/p7U+0Xt/0iPAO1rrN0KYbUaZrvAFGINTlHTpray8/6aS3QjhjzUY9/yW\nfNfGYHThy/E+vw/j73IdcLJSKtL7/dcVc8u4ocAXJQ+UUjMwWhyHa613eheHdWaM936i9/dhGNnC\nKXMmf/W2ycK45WGDUmqYd9l4YDXhnXkX0NNblp8LbNVa3wh8Rx1mltFWq/YfYLRS6ivv44uVMaJf\nnNZ6rlLqDmAFxlWsZVrrJd6T7HLbhD424Dt7KuWbs8Fo9n5NKfUlxpf4xSa2KHmgdATFksw3Aksx\nLni8orXep5Sq8DrNiVtOuewYFa6ZGAXQcqUUwNOE1/sNlb/n4fwZF8Hnq9y4BHjH+xn4Smu92KSc\n/vyNPaO1/sjMXN737CVgjVKqCOOE9bUQ5rkDo5vSbKVUyb2PczG69c49Zt1QD5izAPi3UmoVxpX+\nvwGTlVJVlvchzicaCK31IqXUUKXUOozP01VAPvCeUqpk5NJZWuv9SqlngS+9692htS4yLbjRIl8y\nSr4NY8CcXcACbxm3Umt9b7hm9roJeFkpdSVGC950rfWRMMr8FPCqUmo1xki8f+evQSQjMLo1f6CN\n0VbDNrPWuuQe/9L76rXWf9ZlZovHE+rvDCGEEEIIIYQQ9Y10WxVCCCGEEEII4ZNUHoUQQgghhBBC\n+CSVRyGEEEIIIYQQPknlUQghhBBCCCGETzLaqhCNhFJqIPCId0jnqtb5L5ACFAP5WuuJVa0rhBDh\nyp/yzrteDMbUT7dprZeGJJwQQtRjUnkUohFQSt0KzAByfazaUWvdPQSRhBCiTtSgvAP4J+XnmRRC\nCFENqTyKkFJKvQas0Fq/Xs06G7TWfUKXqlHYDpwJvAmglDoBY94oC3AIY568aCBJKbUQY9LkR7TW\ni8yJK0RwKaWGA3f7aokKV0qpy4BsrfV7ZmepB3yWd1rrbKXUzRgTmQshgkAp1Qb4DMgBXgeKtdYv\nmZtKBJvc8yhCzYOPK7xScQw+rfUCwFlm0VzgKu+J9GLgVozJuh8HTsc48XpKKZUa6qxCiEoNBiLN\nDlEf+CjvPgVuVUqdgtHT4hWMSqUl9EmFaHCGA//TWvfXWj8vFceGSVoeGymllB14EegONAM0cCPw\nLrAZ6APsB87WWmcppfYB84EhGF/KU7XWO5VSO4GhWuvfy17ZV0oNAx4AYoBk4Fat9Qfew1f7Ja2U\ncmutrUqpe4BWQEegLfCy1vohpVQURlejkzDuzbtfaz0vCG9LY9IVeFEpBUal8WfgT+AlrbUbyFBK\nbQA6AxmmpRQiyJRSnTAqE8lAHnCd1nq9UqoH8CwQB6QBT2itn6uqHKpm/5WVrWcCzYGPgB3ACcB6\nYCVwkTfLZK31Nm+Z+gYwFogFLgCaAKcCI5RSe7XWnwfn3Wg0ji3vfsHobdFWKbUC6AL0UUrt01pv\nMi+mEMHnPTf7P+/DDsAHwBHgDIzzsQlAX+BejL+P34BZWutMpdTZGOeG0d6fS7XWXyqlVgLfAicD\nqcC1wD7gfiBOKfUixjkFWut7lVLTvRk8wHfALCACoyzuidF1/HGt9Zt19kaIoJGWx8a2BKbqAAAg\nAElEQVRrEFCotR6McVIUjVGA9MQ4aToBOAyc512/GbBMa90XWA1c411eVSviNcAlWut+wKXA7ABz\nngCMBgYCtyulEjEKqRitdRdgFDDbe8Im/LcNON97Jf5WYBHGezkfQCkVB/QAfjItoRDBVVJWvQU8\nrbXuBdwAfKCUigAuwbgQNQAYCTxYZttjy6GEao5TVdlasp/7AAX0B9p613sXuKxMzoNa64HAv4A7\ntNbLgI+Bu6TiGJBjy7tPtNbnaa2HeJctAW6RiqNowAZgXKjqDlwJHNBa9wc2eR8/DIzxnuN9BvxD\nKWUBLgcmaq17A/8AbvHuzwM4vOXXDcADWuuNGOd6/9VaX1mynlKqFfAkMFpr3QOwAROBe4AM7/nm\nSOAebxdzEebkhLuR8l45OqSUuhrjqmsnjCvuB7wFAMAWjCviJZaUWT7UxyFmAKcqpaYC6RhX0AOx\nXGvtxGgJywQSvcd+yfs69mNUcoR/Sk6grwTe9Fa6PRj3AG1XSo1VSn2DcRXw71rrTLOCClEH4oF2\nWuuPALTW33rLlc7ATcB4pdTtQC/Kl1mVlUPZlR2girK1ZF9/lpSvSqk/gC+8y38H2pfZTUlZ+yNG\nq2UJ6VpZM1WWd+ZFEsIUW7TWewCUUgf5q+zZBUwC2gArva3zNuCQ1tqjlJoMnKaMJ4ZRvjt42XKq\niff3Y7uAWzDOAb/SWu8F0Fpf4M1xJ96/Ra31Ie9o78Mxer+JMCaVx0ZKKXUaRheFp4FXMaZn2AUU\nllnNQ5lCQGtdVOY5SyXrOMo8vwajcFrp/fedAGJ6gKOV5Ckum0sp1RHYpbUuDuAYjYbWeifGfVNo\nrb8HKgwcorW+IcSxhAilyu5ts2CUXfMxBlNZCLwHnON9vqpyqFJVlK0l6xcds7qTypWUw8ceS0YE\n9ZM/5V2ZdS8OUSwhzFJV2WPBqCyu0VqfDqCUigTivT2Q1mMMfLMS2Mhfvc6g6nLqWOXOzZRSTb3r\nW4/ZzurNIsKcdFttvE4B5nlHPd2P0Zp37B+tP4MIHOSvlr+SgicZ42r73VrrJRj37pTsuyZXzqta\ndzUw1XusNIxCLaIG+xVCNE7ZwA7v1XSUUukYXfK3YHTbvltrvRDj6jdKqWNPbvzhT9kaCCflL9AJ\nIURteTDuXRzkvR8cjK6nj2Gcx7kwurSuxOh+X5OyrKTs/A4YqJRq5n38NHAasBzjdoGSCuXp3uOI\nMCeVx8ZrLjBNKfU98CHwDcaV2bJXtsuOjFrV8ruBZ5RS64AswKO1zgJeBn707j8ViPZOxuxztNVj\njnnsuh7gBSBPKbUR+By4Rmud5/slCyEaOQ9Gl/rrlFKbMAbIOdPba+EeYI1S6n/AGIxBI9rjX5lV\nVmVlq6/9VPVc2eXLgDuUUmdWsp4QQlTFVxm2D6P76DxvudgbY5CcjcAPGPcM/w9j+o021Rzj2GN5\nMM4J9wF/A5YqpTZjDFT2Ksb93028x1yFcd/kDwG9QhFSFo9HesEIIYQQQgghhKie3PMoQk4pFQ18\nXcXTd2mtPwllHiGEqAml1MkYrZaVmeC90i6EEEI0ONLyKIQQQgghhBDCJ7nnUQghhBBCCCGET1J5\nFEIIIYQQQgjhk1QehRBCCCGEEEL4JJVHIYQQQgghhBA+SeVRCCGEEEIIIYRPUnkUQgghhBBCCOGT\nVB6FEEIIIYQQQvgklUchhBBCCCGEED5J5VEIIYQQQgghhE9SeRRCCCGEEEII4ZNUHoUQQgghhBBC\n+CSVRyGEEEIIIYQQPknlUQghhBBCCCGET1J5FEIIIYQQQgjhk93sAL5kZOR4arN9cnIMWVn5wYoT\ncpLffPX9NdQ2f2pqvCWIcUQValvW1ZVw/fyHY65wzAThmSscM4GUd6ESivIuXD9jNSWvI7w0lNcB\ngZd3Db7l0W63mR2hViS/+er7a6jv+YW5wvXzE465wjEThGeu/2fvvsOjKtM3jn8nHUggIQm9E3hF\nsAEqqDSxga5i773r7qq7v3VX3UVde1u7axe7KwoioogIYlmxYKEIL6GjAgZIIJ2U8/vjDGwoyUzK\nzJnJ3J/rykWmnfOcMHPPeU55TyTWJM1Lc3mPaTkiS3NZjsZo9s2jiIiIiIiINJ6aRxEREREREQlI\nzaOIiIiIiIgEpOZRREREREREAlLzKCIiIiIiIgGpeRQREREREZGA1DyKiIiIiIhIQGoeRURERERE\nJCA1jyIiIiIiIhKQmkcREREREREJSM2jiIiIiIiIBKTmUURERERERAJS8ygiIiIiIiIBqXkUERER\nERGRgBK8LkBERESap8rqStYWrmF5QS7LC5a5P1uWU1y1lW6tepCT3pc+GX3JSe9Dr/QcWiW28rpk\nERGpQ9ibR2NMIvA80B1IBm631k4Ndx0iIqGmvJNY4DgOeaV5uzSIy1ien8uqrSupqK6gTXI6Oek5\n9E7vw2GdhtElqwPzf17E97/N482lb7B262ocHLqkdiUnow990vuSk+E2ln3S+9KuZXt8Pp/Xiyq1\nUNaJxA4v9jyeDeRZa881xmQAPwAKGBFpjpR30mwUVRSxYntzuOMnl+VbllO4bSvJ8cn0bNOLXm1y\n6JtuGNPjWHql55CT3oe2KW13av6ys9PI61G443ZpZSkrCpazrGApuflLWVawlLk/fcnyglxKKktI\nS2pNn/Q+bkPpbyxz0vvQs00vkuKTvPhzyM6UdSIxwovmcSLwlv/3OKDSgxpERMJBeSdRpaKqgrWF\nq1lWkMvyguU1GsRlrC9ehw8fXdK60qtNb3Iy+nCaOXNHg9g5tQvxcfENmm+LhBb0zxpA/6wBO91f\n7VTza9EvOxrK3PylzPl5Ns8ueIoNJeuJ98XTvXUP/6GvfWv824eMlLZN8SeR4CjrRGKEz3EcT2Zs\njEkDpgBPW2vfqO15lZVVTkJCw76MRKRJ6FixRgom75R1Ei6O47C+aD12k2XppqUs3bR0x+8r8ldQ\nWV1JZotMTJahb2Zf+rbtS9/MvpgsQ++M3rRIbOH1IgCwpWwLSzctZcnGJe7PJvff3E25VFRXkN0y\nm72y9mKvrL0wmWbH7z3Se9TV5CrvGkHrdiJRpUF550nzaIzpCkwCHrfWTqjruXl5hY0qMDs7jby8\nwsBPjFCq33vRvgyNrT87O00rU40QbN41NutCJVLf/5FYV6TVVO1U89OmRayrXMX3axbsOMR0ecEy\niiuKSIlPoad/D2LvNjn0Su/tH7imN21TMkNaWyj/VpXVlazZuorcglx3j2X+UnILlpKbbykoLyA5\nPtm/57TvTofC9s7oQ89OHZV3DRTOdbtgRNrnsaG0HJGluSwHNHz9zosBc9oDM4CrrLWzwz1/EZFw\nUd5JuG0s3cictbOYtWYms9d+zKbSjfRI70HP1r3p3SaHgzoOJSe9D73Tc+iU2pk4X/O7YldCXAK9\n0nPolZ7D0T3G7LjfcRw2lW1iWf5SlvkbywUb5zMp9y3WFK6m2qnGuTkit+FEPGWdSOzw4pzHG4E2\nwHhjzHj/fWOstWUe1CIiEkrKOwmpyupK5m34ltlrPmLWmpn8mPcD7Vq25/BuR3DXsPsY3mUkfbt2\nbzZbyhvD5/OR1SKLrBZZDOl0yE6PlVWWsWbrao8qaxaUdSIxIuzNo7X2GuCacM9XRCTclHcSCr8U\n/szstR8za81MPv35E0orSxjS8RCOzzmJhw5/gn5t99ZlLeopJSGFvm2N12VELWWdSOzwYs+jiIiI\nBKmssowvf/2C2Ws/Zvaamdj8JfRo3ZPDux3BE0c8zSGdh5GamOp1mSIiEgPUPIqIiEQQx3FYXrCM\n2WtnMmvNTP776+f4iOOwzsO4YMAljOo2ml5tentdpoiIxCA1jyIiIh4r2lbIZ7986g50s2YmawpX\ns3fmAEZ1Hc1V+/+RgzoOITk+2esyRUQkxql5FBERCbNqp5pFGxfsOHfx6/VzSU1MZUSXw/nz4L8y\nsuvhdEzt5HWZIiIiO1HzKCIiEgabSjcx52f/ZTTWfMzG0jwGth/MqK6j+fuQWzig3aC6Ll4vIiLi\nOTWPIiIiIbDjMhpr3UNRf/jt+x2X0bjjsHsY3nUkbVMyvS5TREQkaEE1j8aYs4G9gTuBk621L4W0\nKhERjyjvpDG2X0Zj9tqPmbN29o7LaBzXexz/GvkYe2f212U0JGIo70SkvgI2j8aYe4AuwEDgXuAi\nY8z+1to/hbo4EZFwUt5JQ6zeuoo3lrzK9NXvsShvEd1b92B0tyN5/IinOVSX0ZAIpbwTkYYIZs/j\n0bjBMs9au8UYcwSwAFC4iEhzo7yToBRXFDN1+Tv8Z8lrfPHrZwxqP5jLBl3GwZnDdRkNiRbKOxGp\nt2Cax6pdbifv4b6I89Gq6Tz54+NUx1WSldSeB0c9SmpSmtdliUhki8q8k/BwHIev1s/ljcWvMGX5\nZFoltuI0cyb3DP8XfdsasrPTyMsr9LpMkWAp70Sk3uKCeM5E4A2grTHmOuAz4PWQVtUIVdVVXDvr\nai6YfjYDsvbllL1PIbdgKRd9eC4VVRVelycikS2q8k7C49eiX3ho3v0MfW0gJ085jq3btvL0kc/z\nw3mLGT/0n/Rta7wuUaQhlHciUm8B9zxaa+82xhwDrAG6AuOtte+FvLIGKqoopLSyhBmnzKF/1gCy\ns9MY3WEsY94ezV/mXMuDox7TYAUiskfRlncSOmWVZXyw8j1eX/IKc9bOpn/WPly8z2Wc1Oc0Mlto\nhFSJfso7EWmIgHsejTGdgcOttf8HPAacboxpH/LKGqhNcjpPHfUC/bMG7LivQ6uOvHbsW7y7/B2e\nnv/EHl/31bq53PfNXeEqU0QiULTlnTQtx3H4fsM8rp9zHfu82JcbPvs/+qT3ZeZpnzHrtM+5dN8r\n1ThKs6G8E5GGCOacx1dxD2sA+AX4FHgZOCpURYVCv8y9efyIp7n4w3Ppl9mf4V1G7nhsQd6PnPne\nyRRVFHJgh4MZ2fVw7woVES81i7yT+vmt5DfeWvof3ljyCrn5Sxnd7UgeHPkYR/U4hqT4JK/LEwkV\n5Z2I1FswzWNba+2TANbacuAZY8xVoS0rNMb0PJY/D/4rp00dh8noR/+sAaQmpjJtxVRO6XsaKQkt\nuPmLm5h12ufEx8V7Xa6IhF+zyTup27aqbXy0+kPeWPIKM1fPoHd6DmfsdQ6n9j2d9q06eF2eSDgo\n70Sk3oJpHkuNMWOtte8D+IdyLmqKmRtjDgbuttaOaorpBeNPg67n6B5j+Wb9VyzNX0JxRTHn9b+Q\n/xv8N4oqCnnTvsYtX/6dkopiWie14eZDbgtXaSLivWaVd7K7RRsX8saSV3hr6X+oqK5kXM7JvHfS\nDAa2G6zz4SXWhCTvlHUizVswzePlwKvGmJf9t9cC5zR2xsaY6/3TaZIVs2D5fD4GZO3DgKx9dnus\nTXI6Nx58M3fMvYUhnQ7llZ9e5Jy9z6N3ep9wligi3mlWeSeuzWWbmLR0Im/Y11iQ9yPDu4zkjmH3\nMqbncbRIaOF1eSJeafK8U9aJNH/BjLb6A9DfGJMJVFhrtzbRvJcBJ+EeXx8xzut/Ief1vxCAk6f8\njmcXPMVdw+73uCoRCYdYy7vmrLK6kk/WfszrS17lw5Xv0ym1M2fsdTYTjnmVLmldvS5PxHMhyjtl\nnUgzF7B5NMYMBG4E2gI+YwyAY61t1Kgy1tpJxpgejZlGqF2675VcOfMSbjjoH7RObuN1OSISYrGc\nd81Fbv5S3ljyKm/a1yncVsjxOeOYePwUhnQ8RIelitQQirxT1ok0f8EctvoS8CSwCHD89zm1P71p\nZWS0JCGhcYPXZGenNeh1Z2Wewi1zb+TdtRO5buh1jaqhMRpaf6SI9voh+pch2usPI8/yrimyLlQi\n9f2zva4tZVv4z6L/8MIPLzD357kM7z6cu4+8i1P2PoXUpFRPaoo0kVhXJNYUY5p93jWX95iWI7J4\nsRyF5YW8vfhtztvvPOJ8Aa+0GFLBNI/F1trHQl5JLfLzSxr1+uzsNPLyChv8+qv2vYY/z7iWD+yH\njMs5ma6tuzMgcwCpSeF54zS2fq9Fe/0Q/cvQ2PqbS9gHybO8a2zWhUqkvv8zs1ox+YdpvL74Fd5f\nOZW2KZmcbs7k4RFP0rNNLwBKtziUEr7aI/VvFYl1RWJNoLwLl3DkXaS+x+pLyxFZvFiOzWWbOOu9\nU6h0qhjb6aQmO4qmoXkXTPP4oTHmj8B0oGz7ndbaNQ2a4+7CthezIc7rfyED2w/mhYXPcM83d7Ku\n6BdM237MPPVTzzt/EWlyMZ130WBZfi5v577JW7lvsKFoA2N7HceLY15nWOcRusSSSP2EMu+UdSJN\nYH3xOk6bOo60pNb859jJEXH6RTDN43m4IbDrcZs9Gztza+0q4JDGTifUBmTtwwMjHwEgv2wzg17e\nh6nL3+GEnJM8rkxEmljM510kWlu4hneWTWJy7lss3Difwe0P4obDbmB0h7G0SU73ujyRaBWSvFPW\niQSnuKKYlPiUWjd8rt66ilPePZ6ebXrxwjGv0iqxVZgr3LNgRlvtEYY6okZGSlsu3+8q7vvmLo7r\ndYK2dIs0I8q7yLGhZANTl01m8rK3+Wb9VwzI2pdxOScz4ZhX6da6e7M5BErEK8o7kfDaVLqJr9fP\nZUt5AZ+s/ZhpK6ayd2Z/Xh77JpXVFdz8xU0szbcUVxZTUbWNgvJ8Du92JE8e+RzJ8clel79DMKOt\ntgXuAXKAU4F7gT9ba/NDXFvEumK/q3l2wVNMXvYWp/Q93etyRKSJKO+8VVCWz7QVU5m07C2++OVT\nerbpxYk5p/DgyMfo29Z4XZ5Is6K8EwmflVtWcMq7x1NSUUxWi2wGZO3LhGNe5dHvH+KYt0axddtW\nDmg3kMv2vZLUpFQS45JokdCCYV1GkBAXzIGi4RNMNc8AM4CDgUJgHfAKcGwI64pobZLTuW7QX/jL\nnOtondSao3qM8bokEWkayrswK6oo4sOV7/POsreZtWYm7Vq2Z1zOydwy9DYGZO0bEed3iDRTyjuR\nEPjs5zm8tvhlOrTqSHbLdjiOw5M/PsbBHYfyxBHPkBSftOO5h3UZwW1fjicnvS/n978oKr7zgmke\ne1prnzLGXGGtLQduMsbMD3Vhke7K/X5PnM/HBdPP5v4RD3NWv3O9LklEGk95FwZllWXMWjOTyblv\nMWP1B7RKTOX43uOYdMI0DuxwkAYjEwkP5Z1IE6p2qnlm/r+59ct/cELvk1i9dRXf/fYtACf3PY1/\nDLl1t9PdkuOTuf2we7wot8GCaR4rjDFttt8wxvQBqkJXUnTw+Xxcsd/vSY5P4fa5N3Ny39Mi6nhk\nEWkQ5V2IVFZX8unPn/DOsreZtmIqPnwc1+t4XhrzBod2HhZxh+WIxADlnUg9VVZXUrStkOKKYhZs\n/JFFGxdSVFHEhpL1zFozk5KKEp47+mXG9Gy+O/CD+ba+GfgE6GaMmQIMBS4KZVHR5My9zuHer+9g\n6vJ3dP6jSPRT3jWhaqear9fNZVLuRKYuf4fSylKO7jGGx0c/zahuo7XBTcRbyjuRepif9wOnvTCO\nzaWbAWid1IYBWfuQnpxBRkoGtx16F6O6jiY9JcPjSkMrmNFWpxtj5gEHAfHAZdbaDSGvLEqkJKRw\nVr/zeH7BM2oeRaKc8q7xHMfhx7zvmZz7NlOWTWJjaR6Hdz+Su4bdz5E9jomYocZFYp3yTiR4uflL\nOX3qiZwx4AzO7nMRKQkpdE7tEpOnWdTaPBpjbq5x0wG2n8G5vzEGa+0/Q1pZFDm//0U8/sPDLMj7\nkX2y9/O6HBGpJ+Vd49nNS5icO5HJy95mzdbVDOsygr8edBNjex2nazGKRBDlnUjwHMfh3eWTueGz\nv3B4tyN5dOyjbNpY7HVZnqprz6MPN1QOAroAE3GPhT8RWBn60qJHt9bdObL70Vw4/RwO7jiUYV1G\nMC7nZFISUrwuTUSCo7xrgFVbVjJl2SQm5b7F4s2LOLjjUC7f72p+12sc2S2zvS5PRPZMeSdSh7yS\nPJ6e/wR282J+KfqF5QW5/OXAG7l8v6tick/jrmptHq21twAYY/4LDLXWlvhvP4h7jLzU8OCox3l3\n+WQWb/qJ2768mX9++Q+GdDwUn89HenIGfTL60K9tf/bL3r/ZHwstEm2Ud8FbX7yOKcsm8c6yt5m3\n4Vv2yz6A08yZjMs5ic5pXbwuT0QCUN6J7GzllhU8Pf8JluXnkhyfzOe/fEqfDMPh3UYztNNhHNvr\nd3Rr3d3rMiNGMAPmZO1yOwlQ97OLrBZZXDTgUgBuP+xupiybxJLNiwHYVLqRKcsmc/emOyipLOaA\ndgM5q995nNr3DFomtvSybBHZmfJuDzaXbeK95e8yOfct/vvr5/TJ6MuJfU7hsdFP0Tu9j9fliUjD\nKO8k5t379Z38a969HNppGId2HkZ5VRnn9r+Qo7ofExXXXPRCMM3jM8C3xphpuCdUHwc8FNKqolxy\nfDKnmTN3u7+quorcgqW8u2wy939zNwvy5nP/SP0pRSKI8s7PcRymr3qf/3z0MjOWz6BTq86MyzmZ\n2w+7h70z++tLVST6Ke8kpr2w8Fke+/4h/nPcZEZ0HeV1OVEjmNFW7zPGzAZG4B4jf6q19seQV9YM\nxcfFs1fbfux1UD8Gth/EJR+ez/iht9I6uU3gF4tIyCnvXN9t+Jab/3sTC/J+5OKBF/PuuOkMan+g\nGkaRZkR5J7Hqt5LfePWnF7nv27t4+sgJahzrKeBZn8aYRKADkAdswh2N67xQF9bcjep6BFkt2zFx\n6RtelyIifrGed2u2rubyGRcydtIR9G6Tw9yzv+eRMY8wuMNBahxFmplYzztpPqqqq+p8vNqppmhb\nIV/++gWXfngB+7+0F5NyJ/L46Kc5rvfxYaqy+QjmsNXXgG7AYtwtU9u9FJKKYkR8XDzn7X0hExY+\nx0UDLtOKmUhkiMm821JewEPzHuCZ+f9maKdD+fjUz+mfNcDrskQktGIy7yR6OI77tnRw2Fi6kbWF\nq6msriIlPpklmxfz1bov+e63edjNi0lNSqNdi3Zs2baFTaUbqXJ2bygT4xI5vveJvHPCBxyojaIN\nFkzzuA/Qz1rrBHym1MtZ/c7l3q/v4P5v7yarRTZFFUVsKSugW+vujOx6uEZ2Egm/mMq7bVXbeGnR\n89z/7d20b9mBl8a+weHdjvC6LBEJj5jKO4kuZZVlXPzhuXy0+sMd98X74on3xbOtehvdWvdgSMeh\nnN//Ivpn7kNxRREbStaTnpxBZotMEnwJ+Hw+HMfB5/ORmphGu5btdKpYEwimeVwMdAR+bYoZGmPi\ngCeAfYFy4BJr7fKmmHa0yWqRxdUH/JHpK98n3hdHalIaaUmteW/FFP5vzjXcPfwB/jrqT16XKRJL\nYiLvHMfh/ZXvcduX4ymqKOLvQ27lzL3OIT4u3uvSRCR8mizvIjXrJHKtK/qVbdXb8OEjKT6J8qpy\n1mxdTWV1JX0zDH/77M8sK8hl6okzaJGQQnpyBp1SO5MQl0BldSUJccG0MBIKwfzlWwHWGLMQKPPf\n51hrD2/gPMcBSdbaQ4wxBwMP+O+LSTccPJ4bDh6/2/0vLHyW+765i98ferkHVYnErGafdzUHw7lq\n/z9y1QF/JDUx1cuSRMQbTZl3EZd1Epkcx+H2ubfw6PcP7vZYy4SWxPniKaoopGtaN6aM+4AuaV13\ne54aR28F89e/0/+vA/hq/N5QhwLTAay1XxljBjdiWs3W2f3O47HvH+Lf3/6b8/uogRQJk2abd2u2\nruaOubfw7vJ3OGOvs3nmqAl0aNXRq3JExHtNmXcRk3USuaqqq7j769t5fuEzTD5hGntn9qfacais\nriA+LoHMlEwAfi5aS0ZyBqlJaR5XLHsSzKU6PjHGHAYMACYAB1lrP23EPFsDW2vcrjLGxFlrq/f0\n5IyMliQkNO5Qquzs6HzzjR/5D278+EaO6n0Um0s3k9M2hy6tu3hdVr1F69+/pmhfhmivP1y8zLum\nyLo9KSgr4M7P7uThrx5mZI+RfHf5d+zbft96TSNS3z+RWFck1gSRWVck1hRLmjjvwr5uF4zm8h6L\n1uWoqq7i9YWvsyJ/BSsLVvJ+7vuUVpQy/ZwPOKzbYbW+rl27/mGssv6i9f+jqQRsHo0x1wInAJ2B\nicDTxpjnrLX3NXCeW4Gaf/VawwUgP7+kgbNxZWenkZdX2KhpeGVsp5O4O/lu9ntyP5LjkymvKqdP\nel9Sk1Iprijm/P4Xcck+V0T0aFHR/PffLtqXobH1x1JIepl3jc26XW2r2saLi57j/m/upkOrjrw0\n5n+D4dTn/RCp7/9IrCsSa4LIrCsSawLlXSPyLqzrdsGI1PdYfUXrcmwpL+CKjy7m+9/msX+7gXRv\n25XbDrmbI7ofRVpS66hcJoje/489aWjeBXPY6gXAwcBca+0mY8yBwNdAQ1emvgB+B0w0xgwB5jdw\nOs1eYnwiP1zxA3l5W2mVmMrKrSv44pfPqKyuZFtVOfd8fSef/TyHbq27Mz/vRzaVbqSsqoyxPY/j\nqv3/SMfUTl4vgki0uYAozzvHcZi2Yiq3zR1PcUUx/xj6Tw2GIyJ7cgFNl3dat2vGqp1qPlk7i4Ub\nFxDni8PHzjstSitLKNxWSEZKBh1bdWLRpoW8t3wKmS2ymHXaF3RK7dysmq5YF0zzWGWtLTfGbL9d\nClQ2Yp6TgSONMV/4b1/YiGk1e6lJqZQmuacg9GrTm15teu947OgeY7np8+tZV7yOI7ofTYdWHah2\nqnlx0fMMfmUfBrU/kAM7HEz31j3omtaNEV1HEeeL82pRRKJBVOedBsMRkXpoyrzTul0zNW3FVO6Y\newu/FP3MwHa7n8rq4NAioQWpiWks3LSAnwvX0DfDcPUB13BWv3NpkdDCg6ollIJpHucYYx4AUo0x\n44DLgFkNnaH/ekJXNvT18j892vTk1WMn7nb/6eYs5m34hi9++YxvN3zN7LUfs683es8AACAASURB\nVLwglwv6X8Kth97hQaUiUSMq806D4YhIAzRZ3mndrvkpqijili/+zhtLXuG6wX/hogGXkpHS1uuy\nJAIE0zz+BbgU+AE4D3gfeDKURUnj+Hw+Bnc4iMEdDtpx37wN33DSlONo17I9/TL78d2GeawvXs/W\nbQW0TcmkfcsOJMYnkRyfROfUrvTNMORk9PFwKUQ8EVV5t6W8gIfmPcAz8//NIZ0PY+apn9E/a4DX\nZYlIdIiqvJOmVVVdxcbSPFZuXcm6ol+odqrx+Xz48GHzl/DCgmfIbJHF+yfPZN/s/b0uVyJIMKOt\nVhljvsK9HlAlMNNa25jDuMQDg9ofyJNHPs+F08+mZUIrBrYfTJfULnRo1YlNpRtZXrCMyupKyqpK\nWVu4lt9KNnB4tyP464E3cUD7QV6XLxIW0ZJ3uw2GM/Z/g+GIiAQjWvJOAvvht++YuXoGAFVOFVXV\nVVQ77nhF2wdV/LlwLSu2LKOgvIDCbYVsLttEtVNNUlwSHVM7Ee+Lx8HBcRwyW2Rxx7B7OaH3SSTG\nJ3q2XBKZghlt9f+Ay4F3gXjgPWPMndba50NdnDStMT2PZcmFK0lLah1w8IwVW5bzr2/vZeykIzhn\n7wsYP/RW0pJa4zgOq7auJL9sMwe0GxTRI72K1Fek592ug+GMH3obZ+x1tgbDEZF6i/S8k+DkleRx\n5nsn0ys9h7SkNOJ98STEJRLvc78Xqp1qHBw6pXbi1PZn0LZFJqmJqWS1yKZdy/Z0bNVJ3yFSL8Ec\ntno5MMhauxXAGHMr8F9A4RKF0lMygnperza9eWz0U1w84DKunX01+73Yj1aJrSiuKKaoopA4Xxz9\nM/fh/P4XkZaURlaLbA7rPFzNpES7iM27eRu+4Zb//p0FeT9y9QHXcOX+f9BgOCLSGBGbdxIcx3G4\n/tPr6J3ehynjPlATKGERTPO4iZ1H3ypi5wvBSjN2QPtBfHTqp3z682yqnGpS4lPol9kfHz6env8E\nT/74GFVOFeuL19E/cx9uPfQODuxwsNdlizRUxOXd6q2ruHPurRoMR0SaWsTlXSxbtWUlJZW7X/8y\nvaoFv20qoKK6gmrH8R9aWs1vJRv49Oc5zF4zk1mnf6HGUcImmOZxBfBfY8zrQBVwIlBojLkZcKy1\n/wxlgeK9pPgkjuh+9G733zTkZm4acjMAG4rXc+83d3HSlOOYdMJ7aiAlWkVM3hWU5fPQdw/w7Pwn\nNRiOiIRCxORdLJq77kvySn6jtLKEVxe/xJe/flHn8xPiEogjDp/PR7wvnjbJ6ezX7gCeOWrCTpdx\nEwm1YJrHpf6fFP/tjwDH/7uOURQA2rfqwAMjH6Z1Umsu+OBsZpzyCZ3Tunhdlkh9eZ5326q2MWHh\nszzw7T0aDEdEQsnzvIsGpZWlvLb4ZYorimmV2IqT+5xa6ylAFVUVzFo7k8WbFuE4Dvnl+awtXMPW\n8i1UOpX0zxzAMT2P5cVFz/Phyvdp36oDjuMwttdxPDzqCTqldsbZ8V/gapfdmoJNZTotSCJGMKOt\n3mKMSQV6AwuAltbaopBXJlHp70NuwW5ezAnvjOG8/hdy+dCLSaaN12WJBMXLvNNgOCISTlq/C+zX\nol+44IOzyCvNo1d6Dj8XruGOubdyUp9TSYxPoLyyHIDyqnLyyzbzQ973lFWWcUD7QcT74mid1Iae\nbXqRnpyBz+dj7q9fcMZ7JzGo/YHMOeNLeqcHviRaUnwSPl95qBdVJGjBjLY6GnjK/9yhwHxjzDnW\n2g9DXZxEn/i4eJ4+egITFj7HpNyJPPL9v5h8/DT2yd6v3tOqqq6iyqkiKT4pBJWK7M7LvDvhnTHM\nz/tBg+GISFjE2vpdRVUFV868hA9WvgdA66TWdEztjA8fW7dtoaKqAgeHsspSSitLSYpPpqJ6G0M6\nHsKbv3uH9JQMqp1qpq14l6nL3yEpPpnk+BR8+Gid3JrurXtwijmdY3ocS8vElnus4Q8HXEtJRQkt\nElpoT6JErWAOW70LGAa8b61dZ4wZCbwONMtwkcZLTUzl9wdcw9X7/5Fbv7mB0987kXfHfUhORuAt\nbNuVV5VzxtSTKKsqZcq46WogJVw8y7tRXUfz9FEvaDAcEQmXmFm/q6qu4uqPL+W7Dd/yxnGTSI5P\nYUt5Pr8W/wq4jWRSfDI+fKQkpNAyoSXlVeVUOVUM7zKShDh3dTnOF8fveo/jd73HNbiW2hpLkWgR\nTPMY5w8VAKy1i4wxToDXiODz+Xh07KP8tnUTh795KCO7jWZEl1F0TetK17Tu5KT32ePFZ6udaq6Z\ndSVri9ZSXV3FHXNv5dZD7/BgCSQGeZZ31w3+SzhmIyKyXbNev3Mch582LWLm6g+ZnPs2eaW/8e6J\n0zW4jEgjBdM8rjXG/A7AGJMOXA2sCWlV0mzE+eJ4bPRTnNXvXN5fMZXXl7zCuqJfySv9jaS4pJ33\nsvh8+HAPLSmtLGHaSTPZXLaZcVPGkBSfREZK2x1PPbjjEAa1PzD8CyTNnfJORGJF1Obd8oJcrpv9\nB7Zuq/3KInmlG9hYspH92x3AiX1O4dS+p9MxtVMYqxRpnoJpHq8AHga64g7rPAu4LJRFSfMS54vj\nsM7DOazz8B33FW7byk+bfmJ98a/4/IO6udcucvD5fOybvT892/QC4P4RD/Mf+9qO11ZWV3Lbl+O5\nZN8r+OuBN5KalBbeBZLmTHknIrEiKvNuzdbVnDzlePZvN5DTzVm1njvYLbsje6ceQNuUzDBXKNK8\nBTPa6gbgjDDUIjEkLak1B3ccEtRzz+p3Lmf1O3en++au+5JrZl3Js/OfZO/MAXRO7YzPF7ejEa2L\nz+fjsM7DOGOvc2iV2KpB9UvzpLwTkVgRrXl39rRTd1zfcE+nvmyXnZ1GXl5hGCsTiQ21No/GmJU1\nbjrsfM0fx1rbK2RViQQwpONQPj/jGxZtWsC3679mY2ke1U5wp2qUV5Xz2PcPc8/XdzCm53EM7zKS\ndi3bA5AYl0hSfBLxPvfyCInxSXRLaE8rJ1MjozVjyjsRiRXRnnd3DruPgzoOqbNxFJHQqWvP4yj/\nv+NxD2eYAFQBZwFNEizGmBOBU6y1ZzfF9CS2JMYnsn+7gezfbmC9X/v3Ibfwwcr3mLF6OrfPvYWt\n27biOA6V1RWUV5XvdpHezqldGNvzOLJaZOPgUOVU4cNHt9bd6ZfZn/6ZA4jzxTXVokn4hTTvlHUi\nEkGiOu+GdRkRismKSJBqbR6ttasAjDH7WWsvqvHQA8aY7xo7Y2PMw8BRwPeNnZZIfSXGJ3J8zokc\nn3Nirc9xHIfyqnISUiuZ+P07fLT6Q5YV5AIQ74unyqli4tI3WLVlJe1atufI7kfTJjl9x+tbJrbk\ngv6XkN0yO+TLI40TyrxT1olIJFHeiUhjBDNgjs8Yc7i1dhaAMWYsUNEE8/4CmAxc3gTTEmlyPp97\nvafstDTO7HcOZ/Y7Z4/P21JewIxV05nz82x+Lly74/4VW5bz7Pwnuf6gm+jeuvvO097DuZlJ8cn0\nzTC0a9leh8h6JxR5p6wTkUikvBORegumebwYeMkY0xH3uPjVwJ7XovfAGHMxcO0ud19grX3Tf0Fa\nkajWJjmdU80ZnGp2Hneg2qlmwqLnePS7BymvKgs4nZKKEkoqS8hIzqBHm560b9Vxx7mXPnwkxiWQ\nkdKWTqmd2b/dQPpmGHw1DpVNS0zTxYcbr8F5p6wTkSijvBORevM5QQ4yYozJwj2RelNTzdwfMJdb\na8+s7TmVlVVOQkJ8U81SJGI5jsOaLWv4Ke8nlucvZ13huh2XLwHYVrWNTaWbWJG/gnnr5lFSUbLT\n65PikxjefTiDOw7e6fzLY3KOYVj3YY0pLeZ2gzZ13inrRKKG8q7x0xuJ8k4kGjQo74LZ8wiAtXZj\nQ2bQWPn5JYGfVIdoH6pZ9XsvnMvQkrYMTj+MwemH1fm8yupK8kp+2+m+1YWr+Xj1DOb/umin+/fr\nsF+j6s/Ojr3raHqRd43NulCJ1M9wJNYViTVBZNYViTWB8i5cwpF3kfoeqy8tR2RpLssBDc+7oJvH\nEHH8PyJSDwlxCXRM7bTTfR1TOzGk49Ddntucgi6KKetEJFYo70SaMU+bR2vtHGCOlzWIiISask5E\nYoXyTqR5q7V5NMbMruN1jrX28BDUIyISdso7EYkVyjsRaYy69jzeWsdjOhxBRJoT5Z2IxArlnYg0\nWK3No7X2k+2/G2MGAq1wR+VJAHqgQxJEpJlQ3olIrFDeiUhjBDzn0RjzEjAUyAR+AvbHvQjs86Et\nTUQkvJR3IhIrlHci0hBxgZ/CcKA/MBG4HDgYSA5lUSIiHlHeiUisUN6JSL0F0zz+aq3dBiwG9rXW\nLgJi70JIIhILlHciEiuUdyJSb8FcquMXY8wNwMfAvcYYcI+PFxFpbpR3IhIrlHciUm/B7Hm8GFhp\nrf0aeBs4A7gypFWJiHhDeScisUJ5JyL1VmvzaIzp4P81HfjSGNMNeBf4I7A8DLWJiISF8k5EYoXy\nTkQao67DVp8DjgU+Zffr/jhAr1AVJSISZso7EYkVyjsRabC6rvN4rP/XgdbazTUfM8b0CGVRIiLh\npLwTkVihvBORxqi1eTTGdMU9rHWaMWZsjYcSgWnAXiGuTUQkLJR3IhIrlHci0hh1Hbb6T2Ak0AmY\nU+P+SuC9ENYkIhJuyjsRiRXKOxFpsLoOW70QwBjzV2vtPeErSUQkvJR3IhIrlHci0hjBXKrjwpBX\nISISGZR3IhIrlHciUm91Hba63SJjzHjgK6AU8AGOtfbTkFYmIhJ+yjsRiRXKOxGpt2Cax0xglP+n\npl1vB2SMaQO8AqQBScCfrLVz6zsdEZEQUd6JSKxQ3olIvQVsHq21I5twftcBH1lrHzHG9AVeBwY1\n4fRFRBpMeScisUJ5JyINEbB5NMYMA/4CtMI9RzIe6Gat7dGA+T0IlPt/T8Q9TEJEJCIo70QkVijv\nRKQhgjls9VngHuB84BFgLPBdoBcZYy4Grt3l7gustfOMMR2Al4FrAk0nI6MlCQnxQZRZu+zstEa9\n3muq33vRvgzRXn8YeZZ3TZF1oRKp759IrCsSa4LIrCsSa4oxzT7vmst7TMsRWZrLcjRUMM1jqbX2\neWNMDyAfuBSYF+hF1trngOd2vd8Ysw/u4Qx/ttZ+Fmg6+fklQZRYu+zsNPLyChs1DS+pfu9F+zI0\ntv4YC0nP8q6xWRcqkfr+j8S6IrEmiMy6IrEmUN7RjPIuUt9j9aXliCzNZTmg4XkXzKU6So0xbQEL\nDAEcoGVDZmaM2RuYCJxprf2wIdMQEQkh5Z2IxArlnYjUWzB7Hv8FvAmcCHwLnEMQW6ZqcSfuKFyP\nGGMACqy1JzZwWiIiTU15JyKxQnknIvUWzGirE40xb1lrHWPMIKAv8ENDZmatHdeQ14mIhIPyTkRi\nhfJORBqi1ubRGNMZeBQ3TD43xvzNWltAECdTi4hEE+WdiMQK5Z2INEZd5zy+ACzBHcY5BXcYZhGR\n5kh5JyKxQnknIg1W12Grnay1NwIYY2YCP4anJBGRsFPeiUisUN6JSIPVtedx2/ZfrLUV/O/iryIi\nzY3yTkRihfJORBqsrubRF7YqRES8pbwTkVihvBORBqvrsNX+xpiVNW53qnHbsdb2CmFdIiLhpLwT\nkVihvBORBqureewbtipERLylvBORWKG8E5EGq7V5tNauCmMdIiKeUd6JSKxQ3olIY9R1zqOIiIiI\niIgIoOZRREREREREgqDmUURERERERAJS8ygiIiIiIiIBqXkUERERERGRgNQ8ioiIiIiISEBqHkVE\nRERERCSgWq/zGArGmFbAa0A6sA0431r7azhrEBEJB+WdiMQK5Z1I7Aj3nsdLgG+stSOAV4Drwzx/\nEZFwUd6JSKxQ3onEiLDuebTWPmyM2d6wdgfywzl/EZFwUd6JSKxQ3onEjpA1j8aYi4Frd7n7Amvt\nPGPMLKA/cFSo5i8iEi7KOxGJFco7kdjmcxzHkxkbYwwwzVqbU9fzKiurnISE+DBVJSJ74PO6gGgX\nTN4p60QigvKukZR3IlGjQXkX7gFz/gb8bK19BSgGKgO9Jj+/pFHzzM5OIy+vsFHT8JLq9160L0Nj\n68/OTmvCamJHffOusVkXKpH6/o/EuiKxJojMuiKxJlDeNVQk5l2kvsfqS8sRWZrLckDD8y6szSPw\nPPCi/5CHeODCMM9fRCRclHciEiuUdyIxItwD5vwGjAnnPEVEvKC8E5FYobwTiR3hvlSHiIiIiIiI\nRCE1jyIiIiIiIhKQmkcREREREREJSM2jiIiIiIiIBBTTzWNRUSH33nsnRUWBh9ytz3Pr+5qGTLsh\nr2vofEI9rXBMt6lFS50SPbp0yWLu3P96XcZOli/PpUePHixfnut1KTspKirklltu0edPRBotXN/n\nTT2fpppeY6fj9eubejpNPa1wTDfcYrx5LOL++++mqKioSZ9b39c0ZNoNeV1D5xPqaYVjuk0tWuqU\n6LFt2zYWLlzodRk7Wb169Y6fSFJUVMStt96qz5+INFq4vs+bej5NNb3GTsfr1zf1dJp6WuGYbrjF\ndPMoIiIiIiIiwVHzKCIiIiIiIgGpeRQREREREZGA1DyKiIiIiIhIQGoeRUREREREJCA1jyIiIiIi\nIhKQz3Ecr2sQERERERGRCKc9jyIiIiIiIhKQmkcREREREREJSM2jiIiIiIiIBKTmUURERERERAJS\n8ygiIiIiIiIBqXkUERERERGRgBK8LqApGGPigCeAfYFy4BJr7fIaj58M/BVwgFettY94UmgdgliG\nA4EHAB+wHjjHWlvuRa17EkT95wL/B2wBJlhrn/ek0ACMMQcDd1trR+1y/5nANUAlsAC4ylobcde5\nqaP+64CLgTz/XZdba5eGuz6JPrW9pzyqJRF4HugOJAO3W2unelsVGGPigWeAvrjfM1dYaxd5W5XL\nGNMOmAeMjpTPvDHmO9zvAoAV1tqLvawHwBhzA/A7IAl4IlK/o2TP9pQNwGJgAlANLASujsTv7T2p\n+bnFrX8CUbYcu36mgE+JsuXwv69exH1fVQGX+v+dQJQsR83vcGNMDnuo3RhzKXAZ7jru7dbaaXVN\ns7nseRwHJFlrDwH+httkATu+1O/C/QAOBa4yxrT1pMq61bUMPuBp4AJr7TBgOu4bOZLUVX8W8E9g\nhP/nbGNMpNWPMeZ63BXA5F3ubwHcBoy01h4GtAGOC3+Fdautfr+BwLnW2lH+n4hYiZTIFuA95YWz\ngTxr7XDgGOAxj+vZ7jig2p8Pfwfu8LgeYMeKz1NAsde1bGeMSQGokUWR0DiOBIb6v79GAF29rUga\nYNdseBx3PeRG/30+4AQP6wvaLp9bH/Avomw5avlMReP/x1gg3lp7KO567J1E0XLs4Tt8t/eSMaYD\n8AfgEOBo4C5jTFJd020uzeOhuA0V1tqvgMHbH7DWVgF7WWsLgWwgHtjmRZEB1LoMuFuzNwF/MsZ8\nArSNwJX/uurvBfxorS3wb535BhgS/hIDWgachPuBqqkMNwTL/LcTgNJwFhak2uoHGATcaIz5zBjz\nt/CWJVGsrveUFyYC4/2/x+FuJfWctXYKcLn/Zg8g37tqdnIf8G9gndeF1LAf0NIY86Ex5mP/VnGv\nHQUsMMa8A0wF3vO4Hqm/XbOhAhhorf3Uf98HwBFeFNYAu35uo3E59vSZGhSFy2GBBP9OnDa4/UM0\nLceu3+F7ei8dCHxhra2w1m71v2bfuibaXJrH1sDWGrer/IdRAmCtrTbGnAR8D8wGSsJcXzDqWoYs\n3C0Cj+L+R482xnh+CNku6qo/F+hvjGlnjGmJuxe4ZbgLDMRaO4k9rIxaax1rbR6AMeYPQCtr7cxw\n1xdIbfX7vY67cns4cJgx5tiwFSZRK8B7KuystcXW2iJjTBruyuJNXte0nbW2yhjzIvAI8JrX9Rhj\nLsDdEzPDf1ekbAAoBu6z1h4NXAG8WvP72iPZuBvYTsFfk7flSH3tIRv+zs7ruEW4K/8RrZbPbc3P\nblQsB7t/pl4jOpejGHeD4BLcvcGPEEXLsYfv8Jq1F+LW3pr/nUZQ8/5aeR3YTWUrkFbjdpy1trrm\nE/x/wM64u27PC2NtwaprGTYBy6yrEncP3+BdJ+CxWuu31uYD1wFv4wbId8DGsFfYCMaYOGPM/biN\n78le19MAD1trN1trK4BpwAFeFyTSEMaYrsAs4CVr7Rte11OTtfZ83CNFnvEf7u6lC4EjjTGzgf2B\nF40x7T2uCWAp/ubMWpuL+/3W0dOK3O+jGdbaSv9RPWX+0y0kiuySDa/jnte1XRpQ4Elh9bPb5xa3\nEdsuWpZjt88UOzck0bIc1wHTrbUG9//jJSCxxuPRshzb1fxMtMatfdf19zQCHD3TXJrHL3CPS8YY\nMwSYv/0BY0xrY8wnxpgk/yGTxbgnu0aaWpcBWAGkGmN6+28Pwz3RNZLU9X8Qj7urfBhwOrCX//nR\n5CncDQ8n1jh8NSoYY9oAC40xrfyHXhwOfOtxWSL15m9+ZgDXW2sneFzODsaYc2ocDl6K+wVdXcdL\nQs5aO8JaO9I/0NEPwHnW2g1e1uR3Ef5z4o0xnXBXYLw+rPZz3PPkttfUCreplShRSzZ8b4wZ4f99\nDO6ALRFtT59bYHq0LQe7f6ZaAh9H4XJs5n9H1eXjnrYUde+rGvZU+9fAMGNMsn99sR8BeoxmMdoq\nMBl3S832huRC/+iYqdbaZ4wxrwKfGmMqgB+BV7wqtA6BluFi4DX/yv8X1toPPKt0zwLVv32EvTLg\nfmvtZs8qDcyBHSOspuI2WhfhfshmGWPA3ZP3jmcV1m2n+v1//xtxD9kuB2Zaa6d7WaBEnUgZSe5G\n3K3X440x289vGhMBG3QmAS8YY+bgbpW+JpJGw44wzwETjDGf4b6vLtz1SKFws9ZOM8YMN8Z8jbtR\nPSJH05Y67SkbrgEe8Q/+8RPwllfFNYID/Bn3aIaoWY49faaAVUTZcgAPAs8bYz7FHTX2BtxRcKNt\nObbn2W7vJf9oq48An+H+X91ora1zbBif4ygfRUREREREpG7N5bBVERERERERCSE1jyIiIiIiIhKQ\nmkcREREREREJSM2jiIiIiIiIBNRcRlsVkQCMMQcDd/uHAK/tOVOATKACKLHWHhuu+kREREQksql5\nFIkBxpjrgXOAogBPzbHW9g9DSSIiIiISZdQ8SlgZYyYAs621L9bxnO+ttQeEr6qYsAw4CXgZwBiz\nD/Aw4MO9GPZFQAsg3RgzFUjH3Us5zZtyRQIzxgwCrrDWXtpE07sAGGGtvbCer+sJ3GStvSTA84YC\n/wLSgPnA+dbaiiDncRmw1Vr7RjA5KiISKsaYT4CbrbVzGvDaF4Dx1tq1xphpwCXA0TQge8UbOudR\nws0hwAXH1Tg2PWvtJKCyxl3P4F4IexTwAXA97sXN7wdOwG00HzTGZIe7VpFgWWvnNVXj6NfQCx93\nB3rX9QRjTGvgbeBSa+0A/90X12MehwDJ/t91gWYR8VLAdbk6jMTff1hrj7XWrmvEtMQD2vMYo4wx\nCcC/gf5Ae8ACfwJeBxYABwAbgFOttfnGmHXAROAw3CbkNGvtKmPMKmC4tXaNMWYk7paoUcaYEcDt\nQEsgA7jeWvuWf/a+ALVVW2vjjDG3AJ2BHNyVs2ettXcaY1KAx4FDcc/Nu81a+2YT/FliST/g38YY\ncJvGpcB64ClrbTWQZ4z5HugL5HlWpUgd/JlzC+6Kx9fAMCAb+IO1drp/j+DLQCrwOW6eta9jkjuy\nyRhzKm4mtvD/XGKt/cwY8yfgPKAa+NpaewXwCNDTGPOotfYPtUz7SOC/1tqF/tt/AOL981oPvOuv\nfx3wBPBHoAtwAe5n9HfASH8WAxxrjLkKN7/vsNY+Y4wZDdzj/3vkA2daazfVsbwi0sz5c/Je3IZt\nFe7pKwNw8+ce/9EM84DLrLXzjDHx/ucdgLv+9RCQAmwELrfWLg9inj1wj47o6b99C24ulQGdgGnG\nmOHAd8AIAqwXSmTRnsfYNRQos9YeghsOLYCxwL7AA9bafYAC4Gz/89sDM621A4FPgd/7769ta9Hv\ngYuttYNwD0kY38A698Fd6ToY+Jsxpg3uSldLa+1ewBHAeH8zLMFbApzr3/N4PTAN9285EcAYk4r7\n5bLYswpFAqu59TvRn2fX4W64Ancj04vW2v1xm8ug9qQbY3zA5cCx/tfeA/zFv1L1N2CQ/6faGNMJ\nN5O+raNxBHfPZLExZpIx5kfgVmCL/7F2wFRrbT//7XHW2uG4jfG11tqPcZvL8dbaGbgrWsnW2oOB\nY4E7/K+7CXfl7kBgKjAwmOUVkWavDzAKyAXmWWsH4zZtN9XYyHaG/7mHAz/i5tMbwNX+HHwSdwdD\nQziAY629B/gVGGut3Yz2OEYlNY8xylr7Ge6ep6txt5r3wd06/5u19kf/0xbi7jXcbnqN+9sGmMU5\nwL7GmL/jbr1v1cBSZ1lrK621ecBmoA0wHHjVvxwbrLUDrLWVdU1Edtge1FcCLxtjPgPuBn601k4H\nlhpjvgQ+BG7wh7tINNieT4v4Xz4N538rOy8D5QGm4QBYax3gRGCMMeafwPlAK2ttFfBf4FvgZuBx\na+2vBLfVPBH3vJ7rcLfot8JtRLf7wP/vamCW//c17JzBNeuc4v/9JyDL//u7wDvGmEeBxdbaj4Ko\nS0SaP2utLcTdGH+F/8iiObhHh+2Nm5Mn+Z97JvAKYIDN1tp5/gm8BeT4D8GXGKa9NTHKGHM87pbv\nh4DncS/PsBr3kILtHGqsFFlrt9V4zLeH5yTWePxz4GPgE/+/rzWgTIedV/a2z6uCnQ8vywFWBzvw\nRKyy1q7CPW8Ka+13uFshd33OdWEuS6SpbM+KmplUyv/OrXGMMYE2MvkAjDGtcBvEF3Ez7Ef8R1tY\na8f5L3szFphujDl7z5PazTpgrrV2tX8eE4Grtz+4ywawqiCmV+V/neM/OS9UmgAAHgpJREFU/Bxr\n7UP+Aa+OA+41xrxlrb0zyPpEpPkq9f8bB5xtrf0BwBjTDrdBrDTGLDXGjAJGA1fhnrayKx/+w+0D\n2Gn9EUgi8MY7iRLa8xi7RgNv+kfr24C7hX7XQPAReIv6RtzDG8EdaAVjTAbunsyb/Xuzjq4x7foc\n117bcz8FTvPPqx3uyl1SPaYrIs1DoDyZgbvXEGPMCQR/BERf3ObsLtx8GQvEG2OyjDGLgYXW2pv9\n098Hd4NWoI2xM4BBxpgu/tvH4Taowapk5w10uzHGfAWkWWsfxt0wqMHHRKSmWbiNIf5D7ucDXf2P\nvQw8gHuuYhnuWBiZxpjB/uefBqyy1uYHMZ8CIMOfmcnAMTUeC5hlEtnUPMauZ4AzjTHf4Y4A+CXu\nnqiax5/XPJ+otvtvBh42xnyNO0CD4w+WZ4FF/ulnAy2MMS0JboSumvPc9bkO7mASxf7zhj4Cfm+t\nLQ68yCLSzNSWFdt/vxY40n+I1glBTs/B3dP4A+65wfOAQqCbtXYj8BTwjTHmW9xL2kzAPTc43RhT\n66UzrLU/455HOdXfgKbjNqc1661ruWYCNxpjTt7Da7b/fgMwwV/bJbj5LCKxrWaO3Iq7PrYAN1P+\nYq1d6X/sHdwxMF4BsNaWA6cDj/mff5X/dkDW2i3AfcA3uOtpc2s8/B7ugDk9atTWmNFbJcx8jqP/\nKxERaf62j+TsdR0iIiLRSuc8StgZY1rgDjqxJ/+w1r4XznpEJGY4/kOvbtjTY/7RpBvMGNMbeKuW\nhy/2n2ssIhL1lHexS3seRUREREREJCAdviMiIiIiIiIBqXkUERERERGRgNQ8ioiIiIiISEBqHkVE\nRERERCQgNY8iIiIiIiISkJpHERERERERCUjNo4iIiIiIiASk5lFEREREREQCUvMoIiIiIiIiAal5\nFBERERERkYDUPIqIiIiIiEhAah5FREREREQkIDWPIiIiIiIiEpCaRxEREREREQkowesCAsnLK3S8\nrmFXGRktyc8v8bqM3URiXZFYE6iu+sjOTvN5XYOIiIiIeE97HhsgISHe6xL2KBLrisSaQHWJiIiI\niNSXmkcREREREREJSM2jiIiIiIiIBKTmUURERERERAJS8ygiIiIiIiIBqXkUERERERGRgNQ8ioiI\niIiISEBqHkVERERERCQgNY8iIiIiIiISkJpHERERERERCUjNo4iIiIiIiASk5lFEREREREQCUvMo\nIiIiIiIiAal5FBERERERkYDUPIqIiIiIiEhAah5FREREREQkoIRwz9AYkwg8D3QHkoHbrbVTw12H\niIiIiIiIBM+LPY9nA3nW2uHAMcBjHtQgIiIiIiIi9RD2PY/AROAt/+9xQKUHNYiIiIiIiEg9hL15\ntNYWAxhj0nAbyZvCXYOIiIiIiIjUj89xnLDP1BjTFZgEPG6tnVDXcysrq5yEhPiw1CUie+TzugAR\nERER8Z4XA+a0B2YAV1lrZwd6fn5+SeiLqqfs7DTy8gq9LmM3kVhXJNYEqqs+srPTvC5BRERERCKA\nF+c83gi0AcYbY8b77xtjrS3zoBYREREREREJghfnPF4DXBPu+YqIiIiIiEjDeXGpDhEREREREYky\nah5FREREREQkIDWPIiIiIiIiEpCaRxEREREREQlIzaOIiIiIiIgEpOZRREREREREAgrqUh3GmLOB\nvYE7gZOttf/f3p2HyVXWeRu/q7d0lk5IQsuaBQj8HPYQVgXZHEVcAFFwRhRZI+KAigvgjKjXvO6G\nV3TUGRyXDK6MjqgoMBBCZFEkrAZ4EAghJAGaJCTpJL1Vn/mjKhhiuqvS6eqq0PfnuvpKnarT5/nm\npP/ob85znjO7oqkkSZIkSTWl5JXHiPgicCLwdgpl8+yImFXpYJIkSZKk2lHOtNU3Au8BOlJKq4DX\nA2+qaCpJkiRJUk0ppzzmN9kesZn3JEmSJEmvYOWUx2uBnwATIuLDwO+BH1c0lSRJkiSpppRcMCel\n9IWIOAF4GpgEfCql9JuKJ5MkSZIk1YxyFszZBTgupfRR4BvA6RGxQ8WTSZIkSZJqRjnTVn8IPFl8\nvQSYB/xXxRJJkiRJkmpOOeVxQkrp2wAppc6U0tVAa2VjSZIkSZJqSTnlcX1EnLhhIyJeD7QPxuAR\ncVhE3DoYx5IkSZIkVU7JBXOAmcAPI2LDVNXFwBlbO3BEfLx4nEEpopIkSZKkyilntdX7gX0iYiLQ\nnVJaPUhjPw68He+flCRJkqSaV7I8RsRBwOXABCAXEQBZSum4rRk4pfSLiJi6NceQJEmSJA2Ncqat\nzga+DSwAsuJ7Wd+7D67x40fR0FA/VMOVrbW1pdoRNqsWc9ViJjCXJEmStCXKKY9rU0rfqHiSPqxc\nua5aQ/eptbWFtrY11Y7xN2oxVy1mAnNtCcusJEmSoLzyeGNEXATcAHRseDOl9PQgZRiyq5iSJEmS\npIEppzy+l0LB+/Am7++2tYOnlJ4CXrO1x5FeiR5bkfjLi4/1+Xlv1suDbfdz2+I5XHLIpbxx6puG\nMJ0kSZKGm3JWW506BDkkbeTmRTdy1g1nMLF5ewByudxm99tjuz05da/TeM3Orx3KeJIkSRqGyllt\ndQLwRWAa8E7gS8AlKaWVFc4mDUvXP/lrZt50Fp88/NNccOAHqx1HkiRJAqCujH2uBu4BJgJrgGXA\nNZUMJQ1HPb09XHbzZZx305l89sjPWxwlSZJUU8opj7ullP4dyKeUOlNKnwQmVTiXNOxcPOcDzH5w\nNj9/2685e9/zqh1HkiRJeplyFszpjohxGzYiYk8gX7lI0vB0+5J5fP3Eqzii1fsXJUmSVHvKKY9X\nAHOByRFxHXAEcHYlQ0nDTUdPB8vWLmWPCXtUO4okSZK0WeWstnpDRMwHDgXqgfNTSs9VPJk0jDyz\nZjEAu223G11rqhxGkiRJ2ow+y2NEXLHRZgZseFbAgRFBSumzFU0mDSOLVi9k/IjxjGseR9sa26Mk\nSZJqT38L5mwoi4cCp1K4z7ELeAuwT4VzScPKU6ufYsrYqdWOIUmSJPWpzyuPKaVPA0TEncARKaV1\nxe0rKdwDKWmQLFr9FFPG7lbtGJIkSVKfynlUx/abbDcB4yuQRRq2nl69iMljp1Q7hiRJktSnclZb\nvRq4JyKup7BgzluA/1/RVNIws2j1Uxw3+fXVjiFJkiT1qeSVx5TSl4H3As8CzwDvTCl9s9LBpOEi\ny7LitNWp1Y4iSZIk9alkeYyIRmBHoA1YTmG11fdWOpg0XKzoWEF79xrLoyRJkmpaOdNWfwRMBh6h\n8MiODWZXJJE0zDy9+inqcnXsMmbXakeRJEmS+lROedwP+LuUUlZyT0lbbNHqp9h1zCQa6xurHUWS\nJEnqUznl8RFgJ2DpYAwYEXXAN4H9gU7g3JTSE4NxbGlb5P2OkiRJ2haU86iO0UCKiLsi4tbi15yt\nGPNkoCml9BrgUuCrW3EsqSIeeuFBrrjjk9y19A4Anlz1BGf+7h+57vFfbPWxu/PdZNlfL+RbHiVJ\nkrQtKOfK4+eKf2ZAbqPXA/Va4AaAlNIfI+LgrTiWNKieX/c8/3TLTOYunsOhOx3O1Q99i7+fcgK3\nLb6VA151IBfefD53L/sDb9njJPJZnidefJxHli+gM99JXa6eulzh/2NamlqY0DyRprrCVNTxzROY\nOHJ7frfwen6WfsRBrzqYK4/9BrlcjvnP3cMpe55azb+2JEmSVFLJ8phSmhsRRwL7At8HDk0pzduK\nMccCqzfazkdEXUqpdyuOqW1I27o2Jo6cSF2ujizLWN21ivpcPc0NI2moK+f/M7ZcR08H1z/5K94w\n9QRamsZudp8FL/yZ9/z2dKaMncpd776X3cftwYNt93Pl/K/w+aO+zLte/W7ue34+F95yPj98ZDaQ\nY8rYqew9cR9GN44hn/WQ7+0hI2PhqjbmP/cnenp7AFjRsZxn1z7LwTscwjeO/3euTT/hqJ8cSle+\ni0N2PIyTp1keJUmSVNtyG0+f25yI+BBwErALcARwB/Cfxec/brGI+Crwh5TStcXtxSmlSX3t39OT\nzxoa6l/aXrZmGc0NzYwfOf6l99Z2reUjN36EfV61DxcddtFAYmmAlq1ZxqJVizh818Nfeu+epffw\nxIonWN25mknjJjFl3BRGNY5iyZolfOa2z3DTEzcxpmkMe4zfg0WrFvFix4sAjGwYyaw3zmLmjJnk\ncjnWdK7h9qdv5+4ld9Pd201dro6RDSMZ0zSGnVp2YueWnRnfPJ6WES3U5+rJihfEsywjI6Ont4cV\n61ew4PkFXDH3Cp5e9TR7TtyTX57+S8aPHM+tC29l/rL5PPDcAyx4fgFL1izh3Onn8m9v/jea6psq\net6yLOOWhbfQOqqVA3Y8oKJjDYJc6V0kSZL0SlfOZZ73AYdRKHzLI+IQ4G5gQOWRQvl8K3BtRBwO\nPNjfzitXrnvp9YqO5Rzz09ewtnstM/f/AMdMOp6O/Hr+5fbL6Miv5wcP/IC17Z2cu//7BxitPK2t\nLbS1reGF9S/wz7d/olB6jvk6uVyOLMtY3rGcHDlyxV+5c+RoqGtgZMOoil1Z2zjXlurKdzF38S3c\ntfRO6nP1NNTV093bw8SR23Pefu+nsb6RmxfdyGW//xgAzfXNjG4cTUe+k4eX/xmA+e/5M5NaJjPv\nmbmc9uuT2bVlMmMax7B07TMvlcMcOd6x1+ncctrtrFi/nIWrnmRSyySmjtuNLIN7nrubS278KNc+\n9HNWda7i/ufvZWTDKGbscDCjGkeT7+2hI99Je9dqnlv3HM+uXUY+y5f8+20/cnvO3vd8zt7vPC7/\n/cfY/1v709XbxQ6jduSgHQ5m/4kHcfq0M3j1hL2Ztt2erFrRSWEtp8o6oOUwgJf9mw3037CSWltb\nqh1BkiRJNaCcK4/zU0ozIuK+lNL0iGgA7ksp7TeQASMix19XWwU4K6X0WF/7t7WtyaBwpeacG9/L\nkvbFvP+AD3Ll/C/zxIuPA/CW3d/GV4+5ijuW3s7ZN5zBETsfSVNdI/ksT3dvYXGSurp6XrvzkVw4\n/WK681188U+f47m1yzh20usZ3TiaB9seYF3P2sJ9avVNrOtZR29vLw11DTTUNRTLYI4cOZpG1rF0\n5fP8/LGfMrllCo+ueJQvvO4rnDztVM6+4Qxufvqmzf5d6nJ1vGrUDowfMaFYNHvpyHeQZRmjG8fQ\n3DCCTS/yNNQ1vDS9c8OVt+aGkTTWNZDPelm+/gVWdxVmAdfX5+js7mJ04xiO2uVoDtrhYDIylq9/\ngfuev5eFq54o5qh/6V68lZ0reOLFx+npzXPkrq+jPldPvreHxvom7nn2bv5u4t6cte95nHfjmVxw\n4AfZe+K+rO9Zz9rutQC8btdjuODmc3nr7idx8YxL+MDN55FlGd/6++8AheKxcMlSOvKdNOTq2a55\nPP35y8rHuOreWew5Pjhyl6PYv/XAPgt3lmWs61lHe3c7WVaY9Zwrnr9cro66XB3jmsa97BEYWZYx\nd/Ec9p28F9tnu5LL1dZFtRotj7V1kiRJklQV5ZTHWRQWyHkb8DHgfOAvKaWLKx/vr+Xxp4/+iE/M\n+wi3nPZ79thuzz73v33JPP6w9E56s0Lxa6xvoo46uvKdzH74e4xuHM267nWMaRrDoTsewW3PzKEr\n38V+2+/P2BHjWNmxgq7ebkY1jKQuV09Pbzc9xfvYCtMhYXRzM82M4pAdD+Osfc/jx49ew7/ccRkH\nvWoGz7Qv5rtvvIYxTWPIsoxcLkdv1ktvlmd152qWrl3Kix0rySiUweaGZgDWdq+ls6fjZX+XjIx8\nlqc3y8iRI5/l6ehZT0dPx0vTOLcf2crYEYV7+FrGNNOxLs8L619g7uI5PLz8IRrqGhnbNJYDWqcT\nE15NLldHb2+ert4usixjQvMEdm2ZzNGTjmVkw8iXjb+sfSnv+s2pPLJiAR8/5HI+esilmz3n337g\nG1zz8A+44dQ57Pv9PfneCT/k2MnHA7VZhsBcW8LyKEmSJCivPNYD5wHHA/XAHODbKaWeysf7a3n8\njwe+yY6jd+Jt004Z8LHau9u5av4smhuauXD6xYyoHzGg42z6C36WZZx1wxk8smIBvzzpt+w0ZucB\nZ9walSgeqztXcdeyO3nDlBP6vEr33NpnOWD2qzl3v5lc9/j/cP97H6G+rr5imQaDucpneZQkSRKU\nUR4BImI6cBzQA9yYUnq00sE22FAea8nmfsHv6e2hp7fnpSuJ1VDN4nHar09m7uI5fODAi/j0a/61\nJjL1x1zlszxKkiQJoK7UDhHxUeBnwM7AbsBvIuLsSgfb1jTUNVS1OFbbO/Y6HYDT4h+qnESSJElS\nJZSz9OdMYEZKaTVARHwGuBP4biWDadty0rS3M6J+BHtP3KfaUSRJkiRVQMkrj8ByCtNVN2gHVlcm\njrZVI+pHcNK0t1c7hiRJkqQKKefK45PAnRHxYyAPnAKsiYgrgCyl9NlKBpQkSZIkVV855fGx4teG\nG/r+l8KjO2DThxJKkiRJkl6RSpbHlNKnI2IMsAfwEDAqpdRe8WSSJEmSpJpRzmqrxwP3A9cBOwAL\nI+KNlQ4mSZIkSaod5SyY83ngKGBlSmkZcAzw5UqGkiRJkiTVlnLKY12xNAKQUlrAX+95lCRJkiQN\nA+UsmLM4It4KEBHbARcCT1c0lSRJkiSpppRz5fH9wLuBSRQe2zEdOL+SoSRJkiRJtaWc1VafA941\nBFkkSZIkSTWqz/IYEQs32sx4+TMds5TS7hVLJUmSJEmqKf1deTy2+OenKExX/T6QB/4RGJTiGBGn\nAO9IKb17MI4nSZIkSaqMPstjSukpgIg4IKV09kYffTUi7t3agSPia8AbgPu29liSJEmSpMoqZ8Gc\nXEQct2EjIk4Eugdh7DuAC3j5dFhJkiRJUg0q51Ed5wCzI2InCkVvEXBGuQNExDnAhzZ5+30ppZ9F\nxDHlHkeSJEmSVD25LMvK2jEitqewUM7ywRq8WB5nppT+oa99enryWUND/WANKWnLOTtAkiRJZV15\nBCCl9EIlg/Rl5cp11Ri2X62tLbS1ral2jL9Ri7lqMROYa0u0trZUO4IkSZJqQDn3PFZSVvySJEmS\nJNWwsq88VkJK6TbgtmpmkCRJkiSV1md5jIhb+/m+LKV0XD+fS5IkSZJeQfq78viZfj5zqqkkSZIk\nDSN9lseU0twNryPiIGA0hVUXG4CpON1UkiRJkoaNkvc8RsRs4AhgIvAwcCBwB/DdykaTJEmSJNWK\nclZbfR2wD3AtMBM4DBhRyVCSJEmSpNpSTnlcmlLqAh4B9k8pLQB88JskSZIkDSPlPKpjSURcBtwC\nfCkioHD/oyRJkiRpmCjnyuM5wMKU0t3Az4F3ARdUNJUkSZIkqab0WR4jYsfiy+2AuyJiMvAr4CLg\niSHIJkmSJEmqEf1NW/1P4M3APP72uY4ZsHulQkmSJEmSakt/z3l8c/HlQSmlFRt/FhFTKxlKkiRJ\nklRb+iyPETGJwrTW6yPixI0+agSuB15d4WySJEmSpBrR37TVzwLHADsDt230fg/wmwpmkiRJkiTV\nmP6mrZ4FEBGfSCl9cegiSZIkSZJqTTmP6jir4ikkSZIkSTWtv2mrGyyIiE8BfwTWAzkgSynNq2gy\nSZIkSVLNKKc8TgSOLX5tbNPtkiJiHHAN0AI0AR9JKf1hS48jSZIkSRpaJctjSumYQRzvw8D/ppSu\nioi9gB8DMwbx+JIkSZKkCihZHiPiKOBjwGgK90jWA5NTSlMHMN6VQGfxdSOFabCSJEmSpBpXzrTV\n7wBfBM4ErgJOBO4t9U0RcQ7woU3efl9KaX5E7Aj8F3DxlsWVJEmSJFVDLsuyfneIiPtTSgdGxGeB\nucWv+Sml6QMZMCL2ozBd9ZKU0o2l9u/pyWcNDfUDGUrS4MhVO4AkSZKqr5wrj+sjYgKQgMOBW4FR\nAxksIvYGrgXemVJ6qJzvWbly3UCGqqjW1hba2tZUO8bfqMVctZgJzLUlWltbqh1BkiRJNaCc8jgL\n+BlwCnAPcAYwf4DjfY7CKqtXRQTAiymlUwZ4LEmSJEnSEClntdVrI+K/U0pZRMwA9gLuH8hgKaWT\nB/J9kiRJkqTq6rM8RsQuwNcplMXbI+LSlNKLlLFYjiRJkiTplaWun8++BzxK4TEdzRQesyFJkiRJ\nGob6m7a6c0rpcoCIuBl4YGgiSZIkSZJqTX9XHrs2vEgpdQOdlY8jSZIkSapF/ZVHn+0mSZIkSQL6\nn7a6T0Qs3Gh75422s5TS7hXMJUmSJEmqIf2Vx72GLIUkSZIkqab1WR5TSk8NYQ5JkiRJUg3r755H\nSZIkSZIAy6MkSZIkqQyWR0mSJElSSZZHSZIkSVJJlkdJkiRJUkmWR0mSJElSSZZHSZIkSVJJfT7n\nsRIiYjTwI2A7oAs4M6W0dCgzSJIkSZK23FBfeTwX+FNK6WjgGuDjQzy+JEmSJGkAhvTKY0rpaxGx\nobBOAVYO5fiSJEmSpIGpWHmMiHOAD23y9vtSSvMjYg6wD/CGSo0vSZIkSRo8uSzLqjJwRARwfUpp\nWn/79fTks4aG+iFKJWkzctUOIEmSpOob6gVzLgWeSSldA6wFekp9z8qV6yqea0u1trbQ1ram2jH+\nRi3mqsVMYK4t0draUu0IkiRJqgFDWh6B7wI/KE5prQfOGuLxJUmSJEkDMNQL5jwPvGkox5QkSZIk\nbb2hflSHJEmSJGkbZHmUJEmSJJVkeZQkSZIklWR5lCRJkiSVNKzKY3v7Gr70pc/R3j74j0IYrGMP\nxnG29hiDmWHNmq0/15X6d6vkz0MtjCdJkiQNpmFWHtv5yle+QHt7e80eezCOs7XHGMwMg1MeK/Pv\nVsmfh1oYT5IkSRpMw6o8SpIkSZIGxvIoSZIkSSrJ8ihJkiRJKsnyKEmSJEkqyfIoSZIkSSrJ8ihJ\nkiRJKimXZVm1M0iSJEmSapxXHiVJkiRJJVkeJUmSJEklWR4lSZIkSSVZHiVJkiRJJVkeJUmSJEkl\nWR4lSZIkSSVZHiVJkiRJJTVUO8C2IiIOA76QUjo2IqYB3wd6gT8DF6aUhvyBmZtkmg78GvhL8eNv\npZR+NsR5GoHvAlOAEcC/Ao9Q5XPVR65ngN8AjxV3q8b5qgeuBvYCMuD9QCdVPF99ZGqiyudKkiRJ\n1Wd5LENEfBw4A2gvvjULuDylNC8ivgWcBPyyyplmALNSSrOGMscm3g20pZTeExHjgQeA+6jyueoj\n12eAr1b5fL0F6E0pHRkRRwOfK75fzfO1aab/R+E/Jap9riRJklRlTlstz+PA24FccfuglNK84uvf\nAa+vgUwzgDdHxG0R8Z2IGFOFTNcCnyq+rgO6qY1ztblcVT9fKaXrgJnFzanASmBGNc/XZjK9SA2c\nK0mSJFWf5bEMKaVfAD0bvZXb6HU7MG5oE2020x+Bj6aUjgaeBK6oQqa1KaX2iGihUNj+mZf/jFXr\nXG2a65PA3VT5fBWz5SPiB8DXgB9SGz9bm2aqiXMlSZKk6rI8DkzvRq9bKFydqbb/SSndV3z9S2B6\nNUJExCRgDjA7pfRjauRcbZLrJ9TI+QJIKZ0JBPAdoHmjj6p2vjbKdDVwU62cK0mSJFWP5XFg7ive\nDwbwJmBefzsPkRsj4pDi6+OBe4Y6QETsANwEfDyl9P3i21U/V33kqoXzdUZEXFrcXA/kgXuqeb42\nk6kX+EW1z5UkSZKqzwVztsyGVS8vAa6OiCbgYeC/qxfppUwXAF+PiG5gGXB+FbJcTmGa5aciYsM9\nhhcDV1X5XG0u14eBK6t8vn4BfC8ibgMaKZyrR6nuz9bmMj1D9X+2JEmSVGW5LBvyJ0xIkiRJkrYx\nTluVJEmSJJVkeZQkSZIklWR5lCRJkiSVZHmUJEmSJJVkeZQkSZIklWR5lCRJkiSVZHlUxUTE9yLi\nzIiYs9F791UzkyRJkqSBsTyq0nLAMRs2UkrTqxdFkiRJ0kDlsiyrdga9gkTELODNwFKgHjgAaAH+\nmFI6IiJ6U0r+p4UkSZK0jfGXeA2aiHgHcCCwN/BOYBpwEUBK6YgqRpMkSZK0lSyPGkxHAz9PKeVT\nSi8Av612IEmSJEmDw/KowZRRmKq6QU+1gkiSJEkaXJZHDaabgXdGRFNEjAdOKL6fj4j6fr5PkiRJ\nUo2zPGrQpJR+BcwF/gxcBywofnQdcH9EjKBwdVKSJEnSNsbVViVJkiRJJXnlUZIkSZJUkuVRkiRJ\nklSS5VGSJEmSVJLlUZIkSZJUkuVRkiRJklSS5VGSJEmSVJLlUZIkSZJU0v8ByvchHIoAbNwAAAAA\nSUVORK5CYII=\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x1114b7650>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from sklearn.ensemble.partial_dependence import plot_partial_dependence\n",
"\n",
"comp_features = [('annual_inc','dti'),'loan_amnt','meanfico','annual_inc', 'inq_last_6mths', 'revol_util', 'dti']\n",
"\n",
"fig, axs = plot_partial_dependence(est, X_train, comp_features,\n",
" feature_names=list(df.columns),\n",
" figsize=(14, 14), n_jobs=4)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Trying out some alternate plots from `seaborn`."
]
},
{
"cell_type": "code",
"execution_count": 56,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"CPU times: user 2min 3s, sys: 1.14 s, total: 2min 5s\n",
"Wall time: 2min 6s\n"
]
},
{
"data": {
"text/plain": [
"<seaborn.axisgrid.JointGrid at 0x110937ed0>"
]
},
"execution_count": 56,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAboAAAG3CAYAAADCaZBbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNYe3YAAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x10d2e70d0>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%%time\n",
"sns.jointplot(y,df.meanfico.values,annot_kws=dict(stat=\"r\"),\n",
" kind=\"kde\", color=\"#4CB391\").set_axis_labels(\"int_rate\", \"mean_fico\")"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.10"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment