Explainable_Federated_Learning/load_law_data.py at main · Krishna-Rani-t/Explainable_Federated_Learning · 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
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
import torch
import pandas as pd
import numpy as np
from sklearn.utils import shuffle
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, StandardScaler
from sklearn.neighbors import NearestNeighbors

def load_law(url):
    data = pd.read_csv(url)
    data['y'] = data['y'].replace({0: 1, 1: 0})
    data['sex'] = data['sex'].replace({0: 1, 1: 0})
    #data = shuffle(data)

    # Encode categorical columns
    categorical_columns = ["decile1b", "decile3", "fulltime", "fam_inc", "sex", "race", "tier"]
    for col in categorical_columns:
        encoder = LabelEncoder()
        data[col] = encoder.fit_transform(data[col])

    # Normalize numerical columns
    numerical_columns = ["lsat", "ugpa", "zfygpa", "zgpa"]
    scaler = StandardScaler()
    data[numerical_columns] = scaler.fit_transform(data[numerical_columns])
    # Split the data into features and labels
    X = data.drop('y', axis=1)
    y = LabelEncoder().fit_transform(data['y'])
    #print("bismillah")
    return X, y

def load_law_income(url, sensitive_feature):
    data = pd.read_csv(url)
    data['y'] = data['y'].replace({0: 1, 1: 0})
    data['sex'] = data['sex'].replace({0: 1, 1: 0})
    #data = shuffle(data)

    # Encode categorical columns
    categorical_columns = ["decile1b", "decile3", "fulltime", "fam_inc", "sex", "race", "tier"]
    for col in categorical_columns:
        encoder = LabelEncoder()
        data[col] = encoder.fit_transform(data[col])

    # Normalize numerical columns
    numerical_columns = ["lsat", "ugpa", "zfygpa", "zgpa"]
    scaler = StandardScaler()
    data[numerical_columns] = scaler.fit_transform(data[numerical_columns])

    train_df, test_df = train_test_split(data, test_size=0.2, random_state=42)
    data = train_df

    mask = (data['fam_inc'] >=1) & (data['fam_inc'] <=2)
    df1 = data[mask]
    mask = (data['fam_inc'] ==3)
    df2 = data[mask]
    mask = (data['fam_inc'] >=4) & (data['fam_inc'] <=5)
    df3 = data[mask]
    df1 = df1.dropna()
    df2 = df2.dropna()
    df3 = df3.dropna()
    # Split the data into features and labels

    X_client1 = df1.drop('y', axis=1)
    y_client1 = LabelEncoder().fit_transform(df1['y'])

    X_client2 = df2.drop('y', axis=1)
    y_client2 = LabelEncoder().fit_transform(df2['y'])

    X_client3 = df3.drop('y', axis=1)
    y_client3 = LabelEncoder().fit_transform(df3['y'])


    s_client1 = X_client1[sensitive_feature]
    #y_potential_client1 = find_potential_outcomes(X_client1,y_client1, sensitive_feature)
    y_potential_client1 = y_client1
    X_client1 = torch.tensor(X_client1.values, dtype=torch.float32)
    y_client1 = torch.tensor(y_client1, dtype=torch.float32)
    s_client1 = torch.from_numpy(s_client1.values).float()
    y_potential_client1 = torch.tensor(y_potential_client1, dtype=torch.float32)


    s_client2 = X_client2[sensitive_feature]
    #y_potential_client2 = find_potential_outcomes(X_client2,y_client2, sensitive_feature)
    y_potential_client2 = y_client2
    X_client2 = torch.tensor(X_client2.values, dtype=torch.float32)
    y_client2 = torch.tensor(y_client2, dtype=torch.float32)
    s_client2 = torch.from_numpy(s_client2.values).float()
    y_potential_client2 = torch.tensor(y_potential_client2, dtype=torch.float32)


    s_client3 = X_client3[sensitive_feature]
    #y_potential_client3 = find_potential_outcomes(X_client3,y_client3, sensitive_feature)
    y_potential_client3 = y_client3
    X_client3 = torch.tensor(X_client3.values, dtype=torch.float32)
    y_client3 = torch.tensor(y_client3, dtype=torch.float32)
    s_client3 = torch.from_numpy(s_client3.values).float()
    y_potential_client3 = torch.tensor(y_potential_client3, dtype=torch.float32)


    X_test = test_df.drop('y', axis=1)

    y_test = LabelEncoder().fit_transform(test_df['y'])
    sex_column = X_test['sex']
    column_names_list = X_test.columns.tolist()
    #ytest_potential = find_potential_outcomes(X_test,y_test, sensitive_feature)
    ytest_potential = y_test
    ytest_potential = torch.tensor(ytest_potential, dtype=torch.float32)
    X_test = torch.tensor(X_test.values, dtype=torch.float32)
    y_test = torch.tensor(y_test, dtype=torch.float32)


    sex_list = sex_column.tolist()
    data_dict = {}
    data_dict["client_1"] = {"X": X_client1, "y": y_client1, "s": s_client1, "y_pot": y_potential_client1}
    data_dict["client_2"] = {"X": X_client2, "y": y_client2, "s": s_client2, "y_pot": y_potential_client2}
    data_dict["client_3"] = {"X": X_client3, "y": y_client3, "s": s_client3, "y_pot": y_potential_client3}
    #print("bismillah")

    return data_dict, X_test, y_test, sex_list, column_names_list,ytest_potential