Skip to content

Instantly share code, notes, and snippets.

@Erlemar

Erlemar/Mean_encoding.ipynb

Created February 27, 2018 14:31
Show Gist options
  • Save Erlemar/c54e509ee92359d2bc25d742e0a5d40c to your computer and use it in GitHub Desktop.
Save Erlemar/c54e509ee92359d2bc25d742e0a5d40c to your computer and use it in GitHub Desktop.
Download ZIP
An example of mean-encoding.
Display the source blob
Display the rendered blob
Raw
Mean_encoding.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Mean encoding\n",
"\n",
"Mean encoding is a technique to replace categorical values with some figures. In this example I use simple means for each category. Advanced techniques use some kind of regularization.\n",
"\n",
"Let's imagine that we have a dataset where we need to predict shop values based on some features. One of features is shop_id. Using one hot encoding may lead to huge amount of features. Mean-encoding creates a single feature instead."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>shop_id</th>\n",
" <th>target</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1</td>\n",
" <td>11</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>23</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2</td>\n",
" <td>56</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>3</td>\n",
" <td>78</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>4</td>\n",
" <td>89</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>5</td>\n",
" <td>45</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>6</td>\n",
" <td>12</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>1</td>\n",
" <td>34</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>5</td>\n",
" <td>12</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>6</td>\n",
" <td>45</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>7</td>\n",
" <td>12</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>8</td>\n",
" <td>12</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>9</td>\n",
" <td>34</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>0</td>\n",
" <td>90</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>4</td>\n",
" <td>55</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" shop_id target\n",
"0 1 11\n",
"1 1 23\n",
"2 2 56\n",
"3 3 78\n",
"4 4 89\n",
"5 5 45\n",
"6 6 12\n",
"7 1 34\n",
"8 5 12\n",
"9 6 45\n",
"10 7 12\n",
"11 8 12\n",
"12 9 34\n",
"13 0 90\n",
"14 4 55"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df = pd.DataFrame({'shop_id': [1, 1, 2, 3, 4, 5, 6, 1, 5, 6, 7, 8, 9, 0, 4],\n",
" 'target': [11, 23, 56, 78, 89, 45, 12, 34, 12, 45, 12, 12, 34, 90, 55]})\n",
"df"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>shop_id</th>\n",
" <th>target</th>\n",
" <th>encoded</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1</td>\n",
" <td>11</td>\n",
" <td>22.666667</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>23</td>\n",
" <td>22.666667</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2</td>\n",
" <td>56</td>\n",
" <td>56.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>3</td>\n",
" <td>78</td>\n",
" <td>78.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>4</td>\n",
" <td>89</td>\n",
" <td>72.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>5</td>\n",
" <td>45</td>\n",
" <td>28.500000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>6</td>\n",
" <td>12</td>\n",
" <td>28.500000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>1</td>\n",
" <td>34</td>\n",
" <td>22.666667</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>5</td>\n",
" <td>12</td>\n",
" <td>28.500000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>9</th>\n",
" <td>6</td>\n",
" <td>45</td>\n",
" <td>28.500000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>10</th>\n",
" <td>7</td>\n",
" <td>12</td>\n",
" <td>12.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>11</th>\n",
" <td>8</td>\n",
" <td>12</td>\n",
" <td>12.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>12</th>\n",
" <td>9</td>\n",
" <td>34</td>\n",
" <td>34.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>13</th>\n",
" <td>0</td>\n",
" <td>90</td>\n",
" <td>90.000000</td>\n",
" </tr>\n",
" <tr>\n",
" <th>14</th>\n",
" <td>4</td>\n",
" <td>55</td>\n",
" <td>72.000000</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" shop_id target encoded\n",
"0 1 11 22.666667\n",
"1 1 23 22.666667\n",
"2 2 56 56.000000\n",
"3 3 78 78.000000\n",
"4 4 89 72.000000\n",
"5 5 45 28.500000\n",
"6 6 12 28.500000\n",
"7 1 34 22.666667\n",
"8 5 12 28.500000\n",
"9 6 45 28.500000\n",
"10 7 12 12.000000\n",
"11 8 12 12.000000\n",
"12 9 34 34.000000\n",
"13 0 90 90.000000\n",
"14 4 55 72.000000"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"shop_mean = df.groupby('shop_id').target.mean()\n",
"df['encoded'] = df['shop_id'].map(shop_mean)\n",
"df"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.6.4"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment