MetricsForIndividualFairness/Adult.py at main · Magalii/MetricsForIndividualFairness · 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
from aif360.sklearn.datasets import fetch_adult
from scipy.misc import dataset_methods


# class containing all the code needed for experiments on the Adult Census dataset
class Adult:
    def __init__(self):
        self.dataset = fetch_adult()
        self.X_train, self.y_train, self.sample_weights = fetch_adult(subset='train')
        self.X_test, self.y_test, self.sample_weights = fetch_adult(subset='test')

        self.reset_ind()
        self.change_y()


    def change_y(self):
        self.y_train = self.y_train.cat.rename_categories({'<=50K': 0, '>50K': 1})
        self.y_test = self.y_test.cat.rename_categories({'<=50K': 0, '>50K': 1})

    def reset_ind(self):
        self.X_train = self.X_train.reset_index(drop=True)
        self.y_train = self.y_train.reset_index(drop=True)
        self.X_test = self.X_test.reset_index(drop=True)
        self.y_test = self.y_test.reset_index(drop=True)

    def get_dataset(self) -> object:
        return self.dataset

    def get_train(self):
        train_df = self.X_train.copy()
        train_df['y'] = self.y_train
        return train_df

    def get_test(self):
        test_df = self.X_test.copy()
        test_df['y'] = self.y_test
        return test_df

    # reduced adult attributes
    def get_reduced(self):
        columns_to_keep = ['age', 'workclass', 'education', 'marital-status', 'occupation', 'relationship', 'sex',
                    'capital-gain', 'capital-loss', 'hours-per-week']
        X_train_df = self.X_train[columns_to_keep]
        X_test_df = self.X_test[columns_to_keep]
        return X_train_df, self.y_train, X_test_df, self.y_test


    def get_numerical_cols(self):
        return ['age', 'education-num', 'capital-gain', 'capital-loss', 'hours-per-week']