resolver.py
3.44 KB
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
# -*- coding: utf-8 -*-
import codecs
import os
import numpy as np
from natsort import natsorted
from keras.models import Model
from keras.layers import Input, Dense, Dropout, Activation, BatchNormalization
from keras.optimizers import SGD, Adam
IN_PATH = os.path.abspath(os.path.join(os.path.dirname(__file__), 'data',
'prepared_text_files'))
OUT_PATH = os.path.abspath(os.path.join(os.path.dirname(__file__), 'data',
'metrics.csv'))
MODEL = os.path.abspath(os.path.join(os.path.dirname(__file__), 'weights_2017_05_10.h5'))
NUMBER_OF_FEATURES = 1126
def main():
resolve_files()
def resolve_files():
metrics_file = codecs.open(OUT_PATH, 'w', 'utf-8')
write_labels(metrics_file)
anno_files = os.listdir(IN_PATH)
anno_files = natsorted(anno_files)
for filename in anno_files:
print (filename)
textname = filename.replace('.csv', '')
text_data_path = os.path.join(IN_PATH, filename)
resolve(textname, text_data_path, metrics_file)
metrics_file.close()
def write_labels(metrics_file):
metrics_file.write('Text\tAccuracy\tPrecision\tRecall\tF1\tPairs\n')
def resolve(textname, text_data_path, metrics_file):
raw_data = open(text_data_path, 'rt')
test_data = np.loadtxt(raw_data, delimiter='\t')
test_set = test_data[:, 0:NUMBER_OF_FEATURES]
test_labels = test_data[:, NUMBER_OF_FEATURES] # last column consists of labels
inputs = Input(shape=(NUMBER_OF_FEATURES,))
output_from_1st_layer = Dense(1000, activation='relu')(inputs)
output_from_1st_layer = Dropout(0.5)(output_from_1st_layer)
output_from_1st_layer = BatchNormalization()(output_from_1st_layer)
output_from_2nd_layer = Dense(500, activation='relu')(output_from_1st_layer)
output_from_2nd_layer = Dropout(0.5)(output_from_2nd_layer)
output_from_2nd_layer = BatchNormalization()(output_from_2nd_layer)
output = Dense(1, activation='sigmoid')(output_from_2nd_layer)
model = Model(inputs, output)
model.compile(optimizer='Adam', loss='binary_crossentropy', metrics=['accuracy'])
model.load_weights(MODEL)
predictions = model.predict(test_set)
calc_metrics(textname, test_set, test_labels, predictions, metrics_file)
def calc_metrics(textname, test_set, test_labels, predictions, metrics_file):
true_positives = 0.0
false_positives = 0.0
true_negatives = 0.0
false_negatives = 0.0
for i in range(len(test_set)):
if (predictions[i] < 0.5 and test_labels[i] == 0): true_negatives += 1
if (predictions[i] < 0.5 and test_labels[i] == 1): false_negatives += 1
if (predictions[i] >= 0.5 and test_labels[i] == 1): true_positives += 1
if (predictions[i] >= 0.5 and test_labels[i] == 0): false_positives += 1
accuracy = (true_positives + true_negatives) / len(test_set)
precision = true_positives / (true_positives + false_positives)
recall = true_positives / (true_positives + false_negatives)
f1 = 2 * (precision * recall) / (precision + recall)
metrics_file.write('%s\t%s\t%s\t%s\t%s\t%s\n' % (textname,
repr(accuracy),
repr(precision),
repr(recall),
repr(f1),
repr(len(test_set))))
if __name__ == '__main__':
main()