32bytes/encoder.py at main · MLFinalProject/32bytes · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
import pandas as pd
import numpy as np
import category_encoders #pip install category_encoders

def backward_difference_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.BackwardDifferenceEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def count_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.CountEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def helmert_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.HelmertEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def leave_one_out_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.LeaveOneOutEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def m_estimate_encode(input_df_x,input_df_y,y_threshold):
	input_df_x.columns = input_df_x.columns + '_encoded'
	y_label = input_df_y.columns[0]
	input_df_y[y_label] = (input_df_y[y_label] - y_threshold) > 0
	enc = category_encoders.MEstimateEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def one_hot_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.OneHotEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def polynomial_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.PolynomialEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def target_encode(input_df_x,input_df_y):
	input_df_x.columns = input_df_x.columns + '_encoded'
	enc = category_encoders.TargetEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)

def weight_of_evidence_encode(input_df_x,input_df_y,y_threshold):
	input_df_x.columns = input_df_x.columns + '_encoded'
	y_label = input_df_y.columns[0]
	input_df_y[y_label] = (input_df_y[y_label] - y_threshold) > 0
	enc = category_encoders.WOEEncoder(cols = input_df_x.columns)
	return enc.fit_transform(input_df_x,input_df_y)