binary_ensemble/utils.py at main · Idsl-group/binary_ensemble · 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
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
import numpy as np
import pandas
from sklearn.impute import KNNImputer
from sklearn.preprocessing import MinMaxScaler
import matplotlib.pyplot as plt
from sklearn.datasets import make_classification
import numpy as np
from sklearn.model_selection import train_test_split

def process_wearable_dataset(minority):
    dat = pandas.read_csv('./data/wearable.csv')
    new_data = dat[['Weight', 'Age',
                    'mean.Temperature_480',
                    'mean.Temperature_60',
                    'mean.Humidity_60',
                    'mean.Humidity_480',
                    'mean.hr_5',
                    'mean.hr_15',
                    'mean.hr_60',
                    'mean.WristT_5',
                    'mean.WristT_15',
                    'mean.WristT_60',
                    'mean.PantT_5',
                    'mean.PantT_60',
                    'Height',
                    'Coffeeintake',
                    'mean.AnkleT_5',
                    'mean.AnkleT_15',
                    'mean.AnkleT_60',]]

    new_data = np.array(new_data)

    imputer = KNNImputer(n_neighbors=10)
    new_data = imputer.fit_transform(new_data)
    scaler = MinMaxScaler()
    new_data = scaler.fit_transform(new_data)

    y = dat['therm_pref']
    y += 1
    X = np.array(new_data, dtype=float)
    y = np.array(y, dtype=int)
    print(X.shape)

    n, d = X.shape
    ind = np.random.choice(range(n), n, replace=False)
    X = X[ind]
    y = y[ind]

    X_train, X_test, y_train, y_test = train_test_split(
        X, y, test_size=0.2)

    counts = np.bincount(y_train)
    if minority > np.min(counts):
        minority = np.min(counts)

    majority = np.bincount(y_train)[1]

    X_sel = X_train[y_train == 0]
    n, _ = X_sel.shape
    size = minority
    ind = np.random.choice(range(n), size=size, replace=False)
    X_new = X_sel[ind]
    y_new = np.zeros(size).astype(int)

    X_sel = X_train[y_train == 1]
    n, _ = X_sel.shape
    size = majority
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_new = np.concatenate((X_new, new))
    new = np.zeros(size).astype(int) + 1
    y_new = np.concatenate((y_new, new))

    X_sel = X_train[y_train == 2]
    n, _ = X_sel.shape
    size = minority
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_new = np.concatenate((X_new, new))
    new = np.zeros(size).astype(int) + 2
    y_new = np.concatenate((y_new, new))

    n, d = X_new.shape
    indices = np.random.choice(range(n), n, replace=False)
    X_new = X_new[indices]
    y_new = y_new[indices]

    counts = np.bincount(y_new)
    mcir = counts[0]/counts[1] * counts[0]/counts[2]

    print(np.bincount(y_new))
    print(np.bincount(y_test))
    return X_new, y_new, X_test, y_test, mcir


def mnist_imbalanced():
    X_train = np.load('./parsed/X_train.npy')
    y_train = np.load('./parsed/y_train.npy')
    X_test = np.load('./parsed/X_test.npy')
    y_test = np.load('./parsed/y_test.npy')

    X_sel = X_train[y_train == 0]
    n, _ = X_sel.shape
    size = 10
    ind = np.random.choice(range(n), size=size, replace=False)
    X_train_new = X_sel[ind]
    y_train_new = np.zeros(size).astype(int)

    X_sel = X_train[y_train == 1]
    n, _ = X_sel.shape
    size = 4000
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_train_new = np.concatenate((X_train_new, new))
    new = np.zeros(size).astype(int) + 1
    y_train_new = np.concatenate((y_train_new, new))

    X_sel = X_train[y_train == 2]
    n, _ = X_sel.shape
    size = 10
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_train_new = np.concatenate((X_train_new, new))
    new = np.zeros(size).astype(int) + 2
    y_train_new = np.concatenate((y_train_new, new))


    X_sel = X_test[y_test == 0]
    n, _ = X_sel.shape
    size = 100
    ind = np.random.choice(range(n), size=size, replace=False)
    X_test_new = X_sel[ind]
    y_test_new = np.zeros(size).astype(int)

    X_sel = X_test[y_test == 1]
    n, _ = X_sel.shape
    size = 100
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_test_new = np.concatenate((X_test_new, new))
    new = np.zeros(size).astype(int) + 1
    y_test_new = np.concatenate((y_test_new, new))

    X_sel = X_test[y_test == 2]
    n, _ = X_sel.shape
    size = 100
    ind = np.random.choice(range(n), size=size, replace=False)
    new = X_sel[ind]
    X_test_new = np.concatenate((X_test_new, new))
    new = np.zeros(size).astype(int) + 2
    y_test_new = np.concatenate((y_test_new, new))

    assert(X_train_new.shape[0] == y_train_new.shape[0])
    assert(X_test_new.shape[0] == y_test_new.shape[0])


    n, _ = X_train_new.shape
    ind = np.random.choice(range(n), n, replace=False)
    X_train_new = X_train_new[ind]
    y_train_new = y_train_new[ind]


    n, _ = X_test_new.shape
    ind = np.random.choice(range(n), n, replace=False)
    X_test_new = X_test_new[ind]
    y_test_new = y_test_new[ind]

    return X_train_new, y_train_new, X_test_new, y_test_new


def imaginary_dataset():
    X, y = make_classification(n_samples=12000, n_features=20, n_informative=10, n_classes=3)
    X_new = X.copy()
    y_new = y.copy()
    X_new = X_new[y == 1]
    y_new = y_new[y==1]
    sel = X[y==0]
    n, d = sel.shape
    ind = np.random.choice(range(n), size=600, replace=False)
    X_new = np.append(X_new, X[ind], axis=0)
    y_new = np.append(y_new, np.zeros(600).astype(int))
    sel = X[y==2]
    n, d = sel.shape
    ind = np.random.choice(range(n), size=600, replace=False)
    X_new = np.append(X_new, X[ind], axis=0)
    y_new = np.append(y_new, np.zeros(600).astype(int) + 2)
    n, d = X_new.shape
    ind = np.random.choice(range(n), n, replace=False)
    X_new = X_new[ind]
    y_new = y_new[ind]
    scaler = MinMaxScaler()
    X_new = scaler.fit_transform(X_new)
    X_train, X_test, y_train, y_test = train_test_split(
        X_new, y_new, test_size=0.2)

    return X_train, y_train, X_test, y_test

imaginary_dataset()