-
Notifications
You must be signed in to change notification settings - Fork 4
/
test_model.py
184 lines (140 loc) · 5.58 KB
/
test_model.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
# support = 0
# deny = 1
# query = 2
# comment = 3
import numpy as np
import sys
import warnings
from sklearn.exceptions import ConvergenceWarning
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import VotingClassifier
from sklearn.svm import SVC
from sklearn.naive_bayes import GaussianNB
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.model_selection import GridSearchCV
from sklearn.linear_model import LogisticRegression
from sklearn.neural_network import MLPClassifier
from sklearn.gaussian_process import GaussianProcessClassifier
from sklearn.gaussian_process.kernels import RBF
from joblib import dump, load
from sklearn.metrics import confusion_matrix, precision_score, f1_score, accuracy_score, recall_score
import json
import glob
from StanceClassifier.training.preprocesstwitter import PreprocessTwitter
from StanceClassifier.features.extract_features import Features
from StanceClassifier.util import Util
def new_features(l_file, dataset, feature_extractor):
print("Extracting features")
featset = []
labels = []
ids = []
for l in l_file.keys():
r_json = dataset[l][0]
st_id = dataset[l][1]
s_json = dataset[st_id][0]
ids.append(l)
featset.append(np.array(feature_extractor.features(s_json, r_json)))
labels.append(build_labels(l_file, l))
print("Done.")
return np.array(featset), np.array(labels), ids
def open_json(f):
with open(f) as json_file:
data = json.load(json_file)
return data
def new_load_folder(folder):
folders = glob.glob(folder + "/*/*")
dataset = {}
for f in folders:
st_id = f.split("/")[-1]
ct_id = st_id
st_file = open(f + "/source-tweet/" + st_id + ".json")
st_json = json.load(st_file)
dataset[st_id] = []
dataset[st_id].append(st_json)
dataset[st_id].append(st_id)
replies = glob.glob(f + "/replies/*.json")
for r in replies:
r_id = r.split("/")[-1].replace(".json","")
r_file = open(r)
r_json = json.load(r_file)
dataset[r_id] = []
dataset[r_id].append(r_json)
dataset[r_id].append(st_id)
print("Done.",len(dataset.keys())," files loaded!")
return dataset
def build_labels(l_file, l):
label = 3.0
if l_file[l] == "support":
label = 0.0
elif l_file[l] == "deny":
label = 1.0
elif l_file[l] == "query":
label = 2.0
elif l_file[l] == "comment":
label = 3.0
return label
def test(clf, X_test):
return clf.predict(X_test)
def evaluation(y_test, y_pred, model):
f_eval = open(model+".eval", "w")
p_macro = precision_score(y_test, y_pred, average='macro')
p_micro = precision_score(y_test, y_pred, average='micro')
p_weighted = precision_score(y_test, y_pred, average='weighted')
r_macro = recall_score(y_test, y_pred, average='macro')
r_micro = recall_score(y_test, y_pred, average='micro')
r_weighted = recall_score(y_test, y_pred, average='weighted')
f1_macro = f1_score(y_test, y_pred, average='macro')
f1_micro = f1_score(y_test, y_pred, average='micro')
f1_weighted = f1_score(y_test, y_pred, average='weighted')
accuracy = accuracy_score(y_test, y_pred)
cm = confusion_matrix(y_test, y_pred)
print("F1_micro: %f" % f1_micro)
print("F1_macro: %f" % f1_macro)
print("F1_weighted: %f" % f1_weighted)
print("Accuracy: %f" % accuracy)
print(cm)
f_eval.write("Precision macro: " + str(p_macro) + "\nPrecision micro: " + str(p_micro) + "\nPrecision weighted: " + str(p_weighted)+"\n\n")
f_eval.write("Recall macro: " + str(r_macro) + "\nRecall micro: " + str(r_micro) + "\nRecall weighted: " + str(r_weighted)+"\n\n")
f_eval.write("F1 macro: " + str(f1_macro) + "\nF1 micro: " + str(f1_micro) + "\nF1 weighted: " + str(f1_weighted)+"\n\n")
f_eval.write("Acurracy: " + str(accuracy) + "\n\n")
f_eval.write("Confusion Matrix:\n" + str(cm) + "\n\n")
f_eval.close()
def save_json(model, y_pred, ids):
_class = {0.0: "support", 1.0: "deny", 2.0: "query", 3.0: "comment"}
data = {"subtaskaenglish": {}}
for i in range(0, len(ids)):
data["subtaskaenglish"][str(ids[i])] = _class[y_pred[i]]
with open(model+'.json', 'w') as json_file:
json.dump(data, json_file)
train_dict = open_json("/export/data/carolina/rumoureval/2017/semeval2017-task8-dataset/traindev/rumoureval-subtaskA-train.json")
test_dict = open_json("/export/data/carolina/rumoureval/2019/eval-key-taskA.json")
dataset_train = new_load_folder("/export/data/carolina/rumoureval/2017/semeval2017-task8-dataset/rumoureval-data/")
dataset_test = new_load_folder("/export/data/carolina/rumoureval/2019/rumoureval-2019-test-data/twitter-en-test-data/")
print(dataset_train)
#Load resources:
print("Loading resources")
util = Util()
resources = util.loadResources('resources.txt')
print("Done. %d resources added!" % len(resources.keys()))
feature_type = "emb_hc"
feature_extractor = Features(resources)
X, y, ids = new_features(train_dict, dataset_train, feature_extractor)
X_test, y_test, ids_test = new_features(test_dict, dataset_test, feature_extractor)
scaler = StandardScaler()
scaler.fit(X)
X_scl = scaler.transform(X)
#X_scl = X
dump(scaler, "models/scaler")
X_test_scl = scaler.transform(X_test)
#X_test_scl = X_test
#choose classifier
model = "lr"
folder = "models/"
name = folder + model + ".model"
clf = load(name)
#predict
y_pred = test(clf, X_test_scl)
#evaluation
evaluation(y_test, y_pred, name)
#save in the RumourEval format
save_json(name, y_pred, ids_test)