-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathPreprocessing.py
254 lines (197 loc) · 8.75 KB
/
Preprocessing.py
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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
import numpy as np
import csv
import random
def preprocess(metrics, recalculate=False, causal=False):
categories, data = clean_data()
if recalculate:
training_data, training_labels, test_data, test_labels = split_data(data, categories, 0.2, causal=causal)
print("Recalculating data...")
else:
try:
training_data = np.load("COMPAS_train_data.npy")
training_labels = np.load("COMPAS_train_labels.npy")
test_data = np.load("COMPAS_test_data.npy")
test_labels = np.load("COMPAS_test_labels.npy")
for i in range(len(training_labels)):
training_labels[i] = int(training_labels[i])
for i in range(len(test_labels)):
test_labels[i] = int(test_labels[i])
data = np.concatenate((training_data, test_data))
print("Loaded training data")
except:
training_data, training_labels, test_data, test_labels = split_data(data, categories, 0.2, causal=causal)
print("Could not locate data...")
used_metrics = metrics
training_data, reduced_categories, training_predictions = reduce_data(categories, training_data, used_metrics)
np.save("COMPAS_train_decile_scores", training_predictions)
test_data, reduced_categories, test_predictions = reduce_data(categories, test_data, used_metrics)
np.save("COMPAS_test_decile_scores", test_predictions)
mappings = determine_mappings(data, used_metrics)
vectorize_data(training_data, reduced_categories, metrics, mappings)
vectorize_data(test_data, reduced_categories, metrics, mappings)
vectorize_labels(training_labels)
vectorize_labels(test_labels)
training_data = np.array(training_data)
test_data = np.array(test_data)
training_labels = np.array(training_labels)
test_labels = np.array(test_labels)
return training_data, training_labels, test_data, test_labels, reduced_categories, mappings
#######################################################################################################################
def metric_vs_recid(metric):
with open("compas-scores-two-years.csv", "r+") as compas_data:
#print("Opened data file")
reader = csv.reader(compas_data)
totals = {}
possible_values = {}
is_recid = 52
index = -1
categories = reader.__next__()
for i in range(len(categories)):
if metric in categories[i]:
index = i
if index == -1:
print("Couldn't find metric: " + metric)
return
row = reader.__next__()
while row is not None:
if row[is_recid] != "-1":
if row[index] in possible_values:
possible_values[row[index]] = int(possible_values[row[index]]) + int(row[is_recid])
totals[row[index]] = int(totals[row[index]]) + 1
else:
possible_values[row[index]] = row[is_recid]
totals[row[index]] = 1
try:
row = reader.__next__()
except:
break
for value in possible_values:
print(str(value) + ": " + str(int(possible_values[value])*100/int(totals[value])))
print("")
#######################################################################################################################
def clean_data():
pos_data = []
neg_data = []
# Reads data from csv into a list of lists
# Throws out any rows with a -1 for recidivism
with open("compas-scores-two-years.csv", "r+") as compas_data:
is_recid = 52
#print("Opened data file")
reader = csv.reader(compas_data)
categories = reader.__next__()
row = reader.__next__()
while True:
if row[is_recid] != "-1":
if row[is_recid] == "0":
neg_data.append(row)
else:
pos_data.append(row)
try:
row = reader.__next__()
except:
break
if len(pos_data) < len(neg_data):
data = pos_data + random.sample(neg_data, len(pos_data))
else:
data = neg_data + random.sample(pos_data, len(neg_data))
random.shuffle(data)
return categories, data
#######################################################################################################################
def split_data(data, categories, percent_test, causal=False):
if causal:
data = enforce_causal_discrimination(data, categories, "race", "Caucasian")
is_recid = 52
sample_size = int(percent_test * len(data))
while True:
training_data = data[:-sample_size]
test_data = data[-sample_size:]
training_labels = []
test_labels = []
for i in range(len(training_data)):
training_labels.append(training_data[i][is_recid])
zeros = 0
ones = 0
for i in range(len(test_data)):
if test_data[i][is_recid] == "0":
zeros += 1
else:
ones += 1
test_labels.append(test_data[i][is_recid])
if zeros == ones:
break
else:
random.shuffle(data)
np.save("COMPAS_train_data", training_data)
np.save("COMPAS_train_labels", training_labels)
np.save("COMPAS_test_data", test_data)
np.save("COMPAS_test_labels", test_labels)
return training_data, training_labels, test_data, test_labels
#######################################################################################################################
def vectorize_data(data, categories, metrics, mappings):
for metric in metrics:
index = -1
for i in range(len(categories)):
if metric in categories[i]:
index = i
break
for i in range(len(data)):
data[i][index] = mappings[metric][data[i][index]]
#######################################################################################################################
def vectorize_labels(labels):
for i in range(len(labels)):
labels[i] = int(labels[i])
#######################################################################################################################
def reduce_data(categories, data, keep_metrics):
metric_indices = []
reduced_categories = []
for metric in keep_metrics:
metric_indices.append(categories.index(metric))
prediction_index = -1
for i in range(len(categories)):
if "decile_score" in categories[i]:
prediction_index = i
predictions = []
reduced_data = []
for i in range(len(data)):
row = []
for index in metric_indices:
row.append(data[i][index])
reduced_data.append(row)
predictions.append(data[i][prediction_index])
for index in metric_indices:
reduced_categories.append(categories[index])
return reduced_data, reduced_categories, predictions
#######################################################################################################################
def determine_mappings(data, keep_metrics):
with open("compas-scores-two-years.csv", "r+") as compas_data:
#print("Opened data file")
mappings = {}
reader = csv.reader(compas_data)
index = -1
categories = reader.__next__()
for metric in keep_metrics:
mappings[metric] = {}
for i in range(len(categories)):
if metric in categories[i]:
index = i
break
if index == -1:
print("Couldn't find metric: " + metric)
return
possible_values = set()
for i in range(len(data)):
possible_values.add(data[i][index])
for i, value in enumerate(sorted(possible_values)):
mappings[metric][value] = i
return mappings
#######################################################################################################################
def enforce_causal_discrimination(data, categories, reference_metric, reference_value):
index = categories.index(reference_metric)
augmented_data = list.copy(data)
# Loop through training data and add an entry for each class besides the reference class
for i, row in enumerate(data):
if row[index] != reference_value:
duplicate = list.copy(row)
duplicate[index] = reference_value
augmented_data.append(duplicate)
return augmented_data