/
models.py
168 lines (121 loc) · 6.25 KB
/
models.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
from keras.models import Sequential
from keras.layers import Dense
from keras.wrappers.scikit_learn import KerasClassifier
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import KFold
import util
import result
import numpy as np
# define level_1 model
def level_1_model(num_dims):
model = Sequential()
model.add(Dense(4, input_dim=num_dims, init='normal', activation='relu'))
model.add(Dense(3, init='normal', activation='sigmoid'))
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
return model
# define level_2 model
def level2_model(num_dims):
model = Sequential()
model.add(Dense(4, input_dim=num_dims, init='normal', activation='relu'))
model.add(Dense(1, init='normal', activation='sigmoid'))
model.compile(loss='mean_squared_error', optimizer='adam', metrics=['accuracy'])
return model
# define level_3 model
def level3_model(num_dims):
model = Sequential()
model.add(Dense(4, input_dim=num_dims, init='normal', activation='relu'))
model.add(Dense(1, init='normal', activation='sigmoid'))
model.compile(loss='mean_squared_error', optimizer='adam', metrics=['accuracy'])
return model
# define level_4 model
def level4_model(num_dims):
model = Sequential()
model.add(Dense(4, input_dim=num_dims, init='normal', activation='relu'))
model.add(Dense(1, init='normal', activation='sigmoid'))
model.compile(loss='mean_squared_error', optimizer='adam', metrics=['accuracy'])
return model
def cross_validate(model, epochs, batch_size, num_dimensions, seed, xtrain_array, ytrain_array):
"""simple cross validation using keras and sklearn wrapper"""
estimator = KerasClassifier(build_fn=model,
nb_epoch=epochs,
batch_size=batch_size,
num_dims=num_dimensions)
kfold = KFold(n_splits=2, shuffle=True, random_state=seed)
# predicts label level 1.
results = cross_val_score(estimator, xtrain_array, ytrain_array, cv=kfold)
print("Train Baseline: %.2f%% (%.2f%%)" % (results.mean()*100, results.std()*100))
def train_model(params, data):
"""Trains model and returns it"""
xtrain_array, ytrain_array, xtest_array, ytest_array = data
model_function = params['model_function']
num_dimensions = params['num_dimensions']
epochs = params['epochs']
batch_size = params['batch_size']
m = model_function.__call__(*[num_dimensions])
m.fit(xtrain_array, ytrain_array, nb_epoch=epochs, batch_size=batch_size, verbose=0)
scores = m.evaluate(xtest_array, ytest_array)
print(" train accuracy: %s: %.2f%%" % (m.metrics_names[1], scores[1] * 100))
return m
def train_model_and_eval(lvl2_label_name, data, true_positives_all, true_negatives_all, params):
"""Will train, evaluate and update the true positive nested_list"""
(xtrain_array, ytrain_array, xtest_array, y_test) = data
"""
models.cross_validate(model=models.level2_model,
epochs=epochs,
batch_size=batch_size,
num_dimensions=num_dimensions,
seed=seed,
xtrain_array=xtrain_array,
ytrain_array=y_train)
"""
# --------------------- EVAL RESULTS ---------------------
m2 = train_model(params=params,
data=(xtrain_array, ytrain_array, xtest_array, y_test))
flatten_test_prediction = util.flatten_list(m2.predict(xtest_array))
scaled_prediction = util.scale_score_by_frequency(flatten_test_prediction, ytrain_array)
result.print_partial_results(lvl2_label_name, y_test, scaled_prediction, true_positives_all, true_negatives_all)
def train_level_simple(data, train_params, true_positives_all, true_negatives_all, label_names):
"""trains models for an entire level"""
# data is unpacked, altered and then repacked.
(x_train, x_test, text_names_train, text_names_test, text_description_train, text_description_test) = data
for prediction_num in range(min(train_params['number_of_predictions'], len(label_names))):
print('training prediction_num: {}'.format(prediction_num))
label_name = label_names[prediction_num]
# add connection between label name and text.
util.add_label_in_text(x_train, label_name, 'name_text', text_names_train)
util.add_label_in_text(x_test, label_name, 'name_text', text_names_test)
util.add_label_in_text(x_train, label_name, 'description_text', text_description_train)
util.add_label_in_text(x_test, label_name, 'description_text', text_description_test)
train_params['num_dimensions'] = x_train.shape[1]
xtrain_array = np.array(x_train.values.tolist())
xtest_array = np.array(x_test.values.tolist())
y_train = train_params['train_labels'].apply(util.get_label, args=(label_name,))
y_test = train_params['test_labels'].apply(util.get_label, args=(label_name,))
ytrain_array = np.array(y_train.values.tolist())
ytest_array = np.array(y_test.values.tolist())
data = (xtrain_array, ytrain_array, xtest_array, ytest_array)
train_model_and_eval(label_name, data, true_positives_all, true_negatives_all, train_params)
def train_level1(xtrain, ytrain, xtest, ytest):
"""Trains a multiclass classifier for level 1"""
epochs = 25
batch_size = 5
num_dimensions = xtrain.shape[1]
xtrain_array = np.array(xtrain.values.tolist())
xtest_array = np.array(xtest.values.tolist())
"""
models.cross_validate(model=models.level_1_model,
epochs=epochs,
batch_size=batch_size,
num_dimensions=num_dimensions,
seed=seed,
xtrain_array=xtrain_array,
ytrain_array=y_train)
"""
# --------------------- RESULTS ---------------------
m1 = level_1_model(num_dimensions)
m1.fit(xtrain_array, ytrain, nb_epoch=epochs, batch_size=batch_size)
# evaluate the model
predicted = m1.predict(xtest_array)
predicted_level1 = util.from_categorical(predicted)
test_level1 = util.from_categorical(ytest)
return test_level1, predicted_level1