forked from simionAndrei/AppliedNLP_PaperRep
-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_tests.py
140 lines (102 loc) · 5.62 KB
/
run_tests.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
from sklearn.ensemble import AdaBoostClassifier, RandomForestClassifier, VotingClassifier
from sklearn.tree import DecisionTreeClassifier
from xgboost import XGBClassifier
from utils.data_preprocess import generate_final_training_dataset, get_train_test_scores
from logger import Logger
from sklearn.metrics import precision_score, recall_score, f1_score, accuracy_score
from sklearn.metrics import roc_auc_score, mean_squared_error
from sklearn.model_selection import train_test_split
import pickle as pkl
import numpy as np
import sklearn
import json
def run_model(model, params, X_train, y_train, sample_weight, logger):
model.set_params(**params)
logger.log("Start training {} with {} ...".format(type(model).__name__, params))
model.fit(X_train, y_train, sample_weight = sample_weight)
logger.log("Finish training", show_time = True)
return model
def evaluate_model(model, X_test, y_true, y_score, logger, data_type):
y_pred = model.predict(X_test)
y_prob = model.predict_proba(X_test)
y_prob = np.array([elem[1] for elem in y_prob])
logger.log("Scores snippet {}".format(y_prob[:15]))
logger.log("Accuracy {}".format(accuracy_score(y_true, y_pred)))
logger.log("Precision {}".format(precision_score(y_true, y_pred)))
logger.log("Recall {}".format(recall_score(y_true, y_pred)))
logger.log("F1 {}".format(f1_score(y_true, y_pred)))
logger.log("MSE {}".format(mean_squared_error(y_score, y_prob)))
logger.log("AUC {}".format(roc_auc_score(y_true, y_prob, max_fpr = 0.3)))
model_filename = type(model).__name__ + "_" + str(CBAIT_SAMP_W)
model_filename += "_" + "{:.4f}".format(mean_squared_error(y_score, y_prob))
model_filename += "_" + logger.get_time_prefix()
model_filename += ".pkl"
logger.log("Saving model to {} ...".format(model_filename))
with open(logger.get_model_file(model_filename, data_type), "wb") as fp:
pkl.dump(model, fp)
def load_params(model_prefix, data_type, logger):
params_filename = logger.config_dict[model_prefix + ("_S" if data_type == "small" else "_L")]
logger.log("Loading params for {} from {} ...".format(model_prefix, params_filename))
with open(logger.get_model_file(params_filename, data_type)) as fp:
best_params = json.load(fp)
return best_params
# "small", "large", "custom"
DATA_TYPE = "large"
CBAIT_SAMP_W = 4
if __name__ == '__main__':
logger = Logger(show = True, html_output = True, config_file = "config.txt")
if DATA_TYPE == "custom":
final_df = generate_final_training_dataset("large", logger)
custom_df = generate_final_training_dataset("custom", logger)
else:
final_df = generate_final_training_dataset(DATA_TYPE, logger)
X_train, y_train, X_test, y_test, y_score = get_train_test_scores(final_df,
test_size = 0.1 if DATA_TYPE == "small" else 0.1)
logger.log("Splitting data in {} train / {} test".format(y_train.shape[0], y_test.shape[0]))
logger.log("Assign {} weight to clickBaits".format(CBAIT_SAMP_W))
samples_weights = [CBAIT_SAMP_W if int(sample) == 1 else 1 for sample in y_train]
if DATA_TYPE == "custom":
X_custom = custom_df.iloc[:, :-2].values
y_custom = custom_df.iloc[:, -2].values
logger.log("Add {} entries from custom data to training".format(y_custom.shape[0]))
X_train = np.concatenate([X_train, X_custom], axis = 0)
y_train = np.concatenate([y_train, y_custom], axis = 0)
logger.log("Assign {} weight to custom data".format(2.0))
samples_weights += [2.0 for _ in y_custom]
dect_model = DecisionTreeClassifier()
best_params = load_params("BEST_DECT", DATA_TYPE if DATA_TYPE == "small" else "large", logger)
dect_model = run_model(dect_model, best_params, X_train, y_train, samples_weights, logger)
evaluate_model(dect_model, X_test, y_test, y_score, logger, DATA_TYPE)
ada_model = AdaBoostClassifier(DecisionTreeClassifier())
best_params = load_params("BEST_ADA", DATA_TYPE if DATA_TYPE == "small" else "large", logger)
ada_model = run_model(ada_model, best_params, X_train, y_train, samples_weights, logger)
evaluate_model(ada_model, X_test, y_test, y_score, logger, DATA_TYPE)
randf_model = RandomForestClassifier()
best_params = load_params("BEST_RANDF",DATA_TYPE if DATA_TYPE == "small" else "large", logger)
best_params['n_jobs'] = -1
randf_model = run_model(randf_model, best_params, X_train, y_train, samples_weights, logger)
evaluate_model(randf_model, X_test, y_test, y_score, logger, DATA_TYPE)
xgb_model = XGBClassifier()
best_params = load_params("BEST_XGB", DATA_TYPE if DATA_TYPE == "small" else "large", logger)
best_params['n_jobs'] = -1
xgb_model = run_model(xgb_model, best_params, X_train, y_train, samples_weights, logger)
evaluate_model(xgb_model, X_test, y_test, y_score, logger, DATA_TYPE)
dect_model = sklearn.base.clone(dect_model)
ada_model = sklearn.base.clone(ada_model)
randf_model = sklearn.base.clone(randf_model)
xgb_model = sklearn.base.clone(xgb_model)
if DATA_TYPE == "small":
ensemble_names = ["ADA", "RANDF"]
ensemble_weights = [0.5, 0.5]
comb_model = VotingClassifier(estimators = [('ADA', ada_model), ('RANDF', randf_model)],
voting='soft', weights = ensemble_weights, n_jobs = -1)
else:
ensemble_names = ["ADA", "XGB"]
ensemble_weights = [0.5, 0.5]
comb_model = VotingClassifier(estimators = [('ADA', ada_model), ('XGB', xgb_model)],
voting='soft', weights = ensemble_weights, n_jobs = -1)
logger.log("Start training average ensemble {} with {} ...".format(
ensemble_names, ensemble_weights))
comb_model.fit(X_train, y_train, samples_weights)
logger.log("Finish training average ensemble", show_time = True)
evaluate_model(comb_model, X_test, y_test, y_score, logger, DATA_TYPE)