WillKoehrsen

def feature_matrix_from_entityset(es_dict, feature_defs):
    """Run deep feature synthesis from an entityset and feature definitions. 
       Saves feature matrix based on partition number.""" 
    
    # Extract the entityset
    es = es_dict['es']
    
    # Calculate the feature matrix and save
    feature_matrix = ft.calculate_feature_matrix(feature_definitions, 
                                                 entityset=es, 

WillKoehrsen

def total_previous_month(numeric, datetime, time):
    """Return total of `numeric` column in the month prior to `time`."""
    df = pd.DataFrame({'value': numeric, 'time': datetime})
    previous_month = time.month - 1
   
    # Handle January
    if previous_month == 0:
        previous_month = 12
        previous_year = time.year - 1
        

WillKoehrsen

import numpy as np

# Hyperparameter grid
param_grid = {
    'boosting_type': ['gbdt', 'goss', 'dart'],
    'num_leaves': list(range(20, 150)),
    'learning_rate': list(np.logspace(np.log10(0.005), np.log10(0.5), base = 10, num = 1000)),
    'subsample_for_bin': list(range(20000, 300000, 20000)),
    'min_child_samples': list(range(20, 500, 5)),
    'reg_alpha': list(np.linspace(0, 1)),

WillKoehrsen

import lightgbm as lgb

def objective(hyperparameters, iteration):
    """Objective function for grid and random search. Returns
       the cross validation score from a set of hyperparameters."""
    
    # Number of estimators will be found using early stopping
    if 'n_estimators' in hyperparameters.keys():
        del hyperparameters['n_estimators']
    

WillKoehrsen

from hyperopt import fmin

MAX_EVALS = 500

# Optimize
best = fmin(fn = objective, space = space, algo = tpe.suggest, 
            max_evals = MAX_EVALS, trials = bayes_trials)

WillKoehrsen

# Read in results and sort with best on top
results = pd.read_csv('gbm_trials.csv')
results.sort_values('loss', ascending = True, inplace = True)

# Extract the ideal number of estimators and hyperparameters
best_bayes_estimators = int(results.iloc[0, 4])
best_bayes_params = results.iloc[0, 1]

# Re-create the best model and train on the training data
best_bayes_model = lgb.LGBMClassifier(**best_bayes_params, 

WillKoehrsen

  # Write to the csv file ('a' means append)
  of_connection = open(out_file, 'a')
  writer = csv.writer(of_connection)
  writer.writerow([loss, params, iteration, n_estimators, run_time])
  of_connection.close()

WillKoehrsen

import csv

# File to save first results
out_file = 'gbm_trials.csv'
of_connection = open(out_file, 'w')
writer = csv.writer(of_connection)

# Write the headers to the file
writer.writerow(['loss', 'params', 'iteration', 'estimators', 'train_time'])
of_connection.close()

WillKoehrsen

# Sample from the full space
example = sample(space)

# Dictionary get method with default
subsample = example['boosting_type'].get('subsample', 1.0)

# Assign top-level keys
example['boosting_type'] = example['boosting_type']['boosting_type']
example['subsample'] = subsample

WillKoehrsen

# boosting type domain 
boosting_type = {'boosting_type': hp.choice('boosting_type', 
                                            [{'boosting_type': 'gbdt', 
                                                  'subsample': hp.uniform('subsample', 0.5, 1)}, 
                                             {'boosting_type': 'dart', 
                                                  'subsample': hp.uniform('subsample', 0.5, 1)},
                                             {'boosting_type': 'goss',
                                                  'subsample': 1.0}])}