-
Notifications
You must be signed in to change notification settings - Fork 1
/
features_style.py
351 lines (297 loc) · 16.6 KB
/
features_style.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
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
from collections import Counter
from pandas import DataFrame
from pymongo import MongoClient
from pyspark import SparkConf, SparkContext
from pyspark.ml import Pipeline
from pyspark.ml.classification import LinearSVC
from pyspark.ml.classification import LogisticRegression
from pyspark.ml.classification import NaiveBayes, NaiveBayesModel
from pyspark.ml.classification import RandomForestClassifier
from pyspark.ml.classification import DecisionTreeClassifier
from pyspark.ml.evaluation import BinaryClassificationEvaluator
from pyspark.mllib.evaluation import BinaryClassificationMetrics
from pyspark.ml.evaluation import MulticlassClassificationEvaluator
from pyspark.ml.feature import CountVectorizer
from pyspark.ml.feature import IDF
from pyspark.ml.feature import IndexToString, StringIndexer, VectorIndexer, VectorSlicer
from pyspark.ml.feature import MinMaxScaler
from pyspark.ml.feature import OneHotEncoderEstimator
from pyspark.ml.feature import StringIndexer
from pyspark.ml.feature import VectorAssembler
from pyspark.ml.linalg import Vectors
from pyspark.ml.feature import ChiSqSelector
from pyspark.ml.stat import Correlation
from pyspark.mllib.util import MLUtils
from pyspark.sql import SparkSession
from pyspark.sql import SQLContext
from sklearn import metrics
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from pyspark.ml.tuning import ParamGridBuilder, CrossValidator
from spark_stratified import StratifiedCrossValidator
#from sklearn.naive_bayes import MultinomialNB
#from sklearn.utils.multiclass import unique_labels
import csv
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import pymongo
import random
import re
import seaborn as sns
import spacy
def ExtractFeatureImp(featureImp, dataset, featuresCol):
list_extract = []
for i in dataset.schema[featuresCol].metadata["ml_attr"]["attrs"]:
list_extract = list_extract + dataset.schema[featuresCol].metadata["ml_attr"]["attrs"][i]
varlist = pd.DataFrame(list_extract)
varlist['score'] = varlist['idx'].apply(lambda x: featureImp[x])
return(varlist.sort_values('score', ascending = False))
if __name__=='__main__':
spark = SparkSession.builder.appName("Features_Style").getOrCreate()
spark.sparkContext.setLogLevel('ERROR')
client = MongoClient('mongodb://localhost:27017/')
db = client.dataset_fake_and_real_news
collection = db.features_style
#print(collection)
df = pd.DataFrame(collection.find())
#print(type(df))
print("Dataframe pandas")
print(df)
df = df.drop("_id", axis=1)
df_spark = spark.createDataFrame(df)
#print("Dataframe spark")
#print(df_spark.show(truncate = False))
#pipeline stages
stages = []
#...
numericCols = ['avg_word_length_text','num_adj_in_text','num_adj_in_title','num_adv_in_text','num_adv_in_title','num_capital_words_in_text','num_capital_words_in_title','num_exclamation_mark_in_text','num_exclamation_mark_in_title','num_noun_in_text','num_noun_in_title','num_propn_in_text','num_propn_in_title','num_punct_in_text','num_punct_in_title','num_question_mark_in_text','num_question_mark_in_title','num_sentences_in_text','num_word_text']
label_stringIdx = StringIndexer(inputCol = 'type', outputCol = 'label')
stages += [label_stringIdx]
assemblerInputs=numericCols
#print("AssemblerInputs ", assemblerInputs)
assembler = VectorAssembler(inputCols=assemblerInputs, outputCol="features")
stages += [assembler]
pipeline = Pipeline(stages = stages)
pipelineModel = pipeline.fit(df_spark)
df_new = pipelineModel.transform(df_spark)
selectedCols = ['label', 'features']
df_new = df_new.select(selectedCols)
df_new.printSchema()
print("[Data set] ")
print(df_new.show(truncate=False))
#Split train e test
train = df_new.sampleBy("label", fractions={1.0: 0.8, 0.0: 0.8}, seed=2)
test = df_new.exceptAll(train)
#...
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
paramGrid = ParamGridBuilder().build()
#Random Forest
classifier = RandomForestClassifier(featuresCol = 'features', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=classifier, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(train)
print("[Random Forest] Default accuracy: ",cvModel.avgMetrics)
#Decision Tree
classifier = DecisionTreeClassifier(featuresCol = 'features', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=classifier, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(train)
print("[Decision Tree] Default accuracy: ",cvModel.avgMetrics)
#Linear SVC
classifier = LinearSVC(featuresCol = 'features', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=classifier, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(train)
print("[Linear SVC] Default accuracy: ",cvModel.avgMetrics)
#Logistic Regression
classifier = LogisticRegression(featuresCol = 'features', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=classifier, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(train)
print("[Logistic Regression] Default accuracy: ",cvModel.avgMetrics)
#Naive Bayes
classifier = NaiveBayes(featuresCol = 'features', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=classifier, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(train)
print("[Naive Bayes] Default accuracy: ",cvModel.avgMetrics)
print("\n")
print("[Decision Tree con Gini Index] Feature Selection - Gini index")
random_fs = RandomForestClassifier(numTrees=100, maxDepth=5, featuresCol = 'features', labelCol = 'label')
model = random_fs.fit(train)
#print(model.featureImportances)
varlist = ExtractFeatureImp(model.featureImportances, train, "features")
list_acc = []
for i in range (1,20):
varidx = [x for x in varlist['idx'][0:i]]
slicer = VectorSlicer(inputCol="features", outputCol="features2", indices=varidx)
new_train = slicer.transform(train)
new_train = new_train.drop('rawPrediction', 'probability', 'prediction')
dt2 = DecisionTreeClassifier(featuresCol = 'features2', labelCol = 'label')
paramGrid = ParamGridBuilder().build()
cv = StratifiedCrossValidator(estimator=dt2, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(new_train)
print(" [Decision Tree con Gini Index] Accuracy con num_features = ",i, cvModel.avgMetrics)
list_acc.append(cvModel.avgMetrics)
best_idx = list_acc.index(max(list_acc))+1
print("[Decision Tree con Gini Index] Il numero di features ottimo è ", best_idx)
print(ExtractFeatureImp(model.featureImportances, train, "features").head(best_idx))
varidx = [x for x in varlist['idx'][0:best_idx]]
slicer = VectorSlicer(inputCol="features", outputCol="features2", indices=varidx)
new_train = slicer.transform(train)
new_train = new_train.drop('rawPrediction', 'probability', 'prediction')
dt2 = DecisionTreeClassifier(featuresCol = 'features2', labelCol = 'label')
#print(dt2.explainParams())
paramGrid = (ParamGridBuilder()
.addGrid(dt2.maxDepth, [2, 5, 10, 15])
.addGrid(dt2.impurity, ['gini', 'entropy'])
.addGrid(dt2.minInstancesPerNode, [1, 2, 3])
.build())
cv = StratifiedCrossValidator(estimator=dt2, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(new_train)
print("[Decision Tree con Gini Index] Selezione degli iperparametri ")
print("[Decision Tree con Gini Index] Modello migliore ", cvModel.bestModel)
best_params = cvModel.getEstimatorParamMaps()[ np.argmax(cvModel.avgMetrics) ]
print("[Decision Tree con Gini Index] Parametri migliori", best_params)
varidx = [x for x in varlist['idx'][0:best_idx]]
slicer = VectorSlicer(inputCol="features", outputCol="features2", indices=varidx)
new_test = slicer.transform(test)
predictions = cvModel.bestModel.transform(new_test)
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
accuracy = evaluator.evaluate(predictions)
evaluator = BinaryClassificationEvaluator()
roc_auc = evaluator.evaluate(predictions)
print ("[Decision Tree con Gini Index - parametri migliori] Accuracy Score: {0:.4f}".format(accuracy))
print ("[Decision Tree con Gini Index - parametri migliori] ROC-AUC: {0:.4f}".format(roc_auc))
label=[1.0,0.0]
predictions_pandas = predictions.toPandas()
lbl = predictions_pandas['label'].tolist()
prd = predictions_pandas['prediction'].tolist()
cm_dt_gini = confusion_matrix(lbl, prd, labels=label)
#print("Confusion matrix ",cm)
tp = cm_dt_gini[0][0]
#print("TP ", tp)
fp = cm_dt_gini[1][0]
#print("FP ", fp)
precision = tp /(tp + fp)
print("[Decision Tree con Gini Index - parametri migliori] Precision: {0:.4f}".format(precision))
print("[Decision Tree con Gini Index - parametri migliori] Matrice di confusione: ", cm_dt_gini)
ax= plt.subplot()
sns.heatmap(cm_dt_gini, annot=True, ax = ax, cmap='Blues', fmt="d"); #annot=True to annotate cells
# labels, title and ticks
ax.set_xlabel('Predicted labels');ax.set_ylabel('True labels');
ax.set_title('Confusion Matrix');
ax.xaxis.set_ticklabels(['real','fake']); ax.yaxis.set_ticklabels(['real','fake'])
plt.savefig('/home/vmadmin/fake_and_real_news_project/confusion_decision_tree_gini.pdf', dpi=300, bbox_inches="tight")
print("[Decision Tree] Feature Selection - ChiSquared Selector")
#...
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
selector = ChiSqSelector(numTopFeatures=10, featuresCol="features", outputCol="selectedFeatures", labelCol="label")
result = selector.fit(train)
#indici=result.selectedFeatures
#print("indici")
#print(indici)
result2=result.transform(train)
colname = df_spark.toPandas().columns[result.selectedFeatures]
#print (colname)
list_acc_2 = []
for i in range (1,20):
selector = ChiSqSelector(numTopFeatures=i, featuresCol="features", outputCol="selectedFeatures", labelCol="label")
paramGrid = ParamGridBuilder().build()
dt3 = DecisionTreeClassifier(featuresCol = 'selectedFeatures', labelCol = 'label')
cv = StratifiedCrossValidator(estimator=dt3, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
result = selector.fit(train)
result2=result.transform(train)
cvModel = cv.fit(result2)
print(" [Decision Tree con ChiSq] Accuracy con numTopFeatures = ",i, cvModel.avgMetrics)
list_acc_2.append(cvModel.avgMetrics)
best_numTopFeatures = list_acc_2.index(max(list_acc_2))+1
print("[Decision Tree con ChiSq] Il numero di features ottimo è ", best_numTopFeatures)
selector = ChiSqSelector(numTopFeatures=best_numTopFeatures, featuresCol="features", outputCol="selectedFeatures", labelCol="label")
new_train = selector.fit(train)
new_train2 = new_train.transform(train)
print("[Decision Tree con ChiSq] Le top features sono: ", df_spark.toPandas().columns[new_train.selectedFeatures])
dt3 = DecisionTreeClassifier(featuresCol = 'selectedFeatures', labelCol = 'label')
#print(dt2.explainParams())
paramGrid = (ParamGridBuilder()
.addGrid(dt3.maxDepth, [2, 5, 10, 15])
.addGrid(dt3.impurity, ['gini', 'entropy'])
.addGrid(dt3.minInstancesPerNode, [1, 2, 3])
.build())
cv = StratifiedCrossValidator(estimator=dt3, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
cvModel = cv.fit(new_train2)
print("[Decision Tree con ChiSq] Selezione degli iperparametri ")
print("[Decision Tree con ChiSq] Modello migliore ", cvModel.bestModel)
best_params = cvModel.getEstimatorParamMaps()[ np.argmax(cvModel.avgMetrics) ]
print("[Decision Tree con ChiSq] Parametri migliori", best_params)
new_test = new_train.transform(test)
predictions = cvModel.bestModel.transform(new_test)
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
accuracy = evaluator.evaluate(predictions)
evaluator = BinaryClassificationEvaluator()
roc_auc = evaluator.evaluate(predictions)
print ("[Decision Tree con ChiSq - parametri migliori] Accuracy Score: {0:.4f}".format(accuracy))
print ("[Decision Tree con ChiSq - parametri migliori] ROC-AUC: {0:.4f}".format(roc_auc))
label=[1.0,0.0]
predictions_pandas = predictions.toPandas()
lbl = predictions_pandas['label'].tolist()
prd = predictions_pandas['prediction'].tolist()
cm_chi = confusion_matrix(lbl, prd, labels=label)
#print("Confusion matrix ",cm)
tp = cm_chi[0][0]
#print("TP ", tp)
fp = cm_chi[1][0]
#print("FP ", fp)
precision = tp /(tp + fp)
print("[Decision Tree con ChiSq - parametri migliori] Precision: {0:.4f}".format(precision))
print("[Decision Tree con ChiSq - parametri migliori] Matrice di confusione: ", cm_chi)
plt.clf()
ax1= plt.subplot()
sns.heatmap(cm_chi, annot=True, ax = ax1, cmap='Blues', fmt="d"); #annot=True to annotate cells
# labels, title and ticks
ax1.set_xlabel('Predicted labels');ax1.set_ylabel('True labels');
ax1.set_title('Confusion Matrix');
ax1.xaxis.set_ticklabels(['real','fake']); ax1.yaxis.set_ticklabels(['real','fake'])
plt.savefig('/home/vmadmin/fake_and_real_news_project/confusion_decision_tree_chisq.pdf', dpi=300, bbox_inches="tight")
print("\n\n\n\n")
print("[Random Forest] Selezione degli iperparametri ")
#...
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
random_forest = RandomForestClassifier(featuresCol = 'features', labelCol = 'label', featureSubsetStrategy='auto', minInstancesPerNode=2)
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
paramGrid = (ParamGridBuilder()
.addGrid(random_forest.maxDepth, [5, 10])
.addGrid(random_forest.numTrees, [20, 50, 100])
.addGrid(random_forest.subsamplingRate, [0.7, 1.0])
.build())
cv = StratifiedCrossValidator(estimator=random_forest, estimatorParamMaps=paramGrid, evaluator=evaluator, numFolds=10)
# Run cross validations. This can take about 6 minutes since it is training over 20 trees!
cvModel = cv.fit(train)
print("[Random Forest] Selezione degli iperparametri ")
print("[Random Forest] Modello migliore ", cvModel.bestModel)
best_params = cvModel.getEstimatorParamMaps()[ np.argmax(cvModel.avgMetrics) ]
print("[Random Forest] Parametri migliori", best_params)
predictions = cvModel.transform(test)
evaluator = MulticlassClassificationEvaluator(labelCol="label", predictionCol="prediction", metricName="accuracy")
accuracy = evaluator.evaluate(predictions)
evaluator = BinaryClassificationEvaluator()
roc_auc = evaluator.evaluate(predictions)
print ("[Random Forest - parametri migliori] Accuracy Score: {0:.4f}".format(accuracy))
print ("[Random Forest - parametri migliori] ROC-AUC: {0:.4f}".format(roc_auc))
label=[1.0,0.0]
predictions_pandas = predictions.toPandas()
lbl = predictions_pandas['label'].tolist()
prd = predictions_pandas['prediction'].tolist()
cm_rand = confusion_matrix(lbl, prd, labels=label)
#print("Confusion matrix ",cm)
tp = cm_rand[0][0]
#print("TP ", tp)
fp = cm_rand[1][0]
#print("FP ", fp)
precision = tp /(tp + fp)
print ("[Random Forest - parametri migliori] Precision: {0:.4f}".format(precision))
print ("[Random Forest - parametri migliori] Matrice di confusione: ", cm_rand)
plt.clf()
ax2= plt.subplot()
sns.heatmap(cm_rand, annot=True, ax = ax2, cmap='Blues', fmt="d"); #annot=True to annotate cells
# labels, title and ticks
ax2.set_xlabel('Predicted labels');ax2.set_ylabel('True labels');
ax2.set_title('Confusion Matrix');
ax2.xaxis.set_ticklabels(['real','fake']); ax2.yaxis.set_ticklabels(['real','fake'])
plt.savefig('/home/vmadmin/fake_and_real_news_project/confusion_random_forest.pdf', dpi=300, bbox_inches="tight")