Experimento 2

embeddings_RNN.py

@@ -142,10 +142,10 @@ def classification_embedings_rnn(tweets_train, tweets_train_labels_numeric, twee
     #number of features:_32 each vector of 200 dim is converted to a vector of 32 dim
     #number of features:_32 each vector of 200 dim is converted to a vector of 32 dim
     model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2))
     model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2))
     #model.add(Bidirectional(LSTM(2,dropout=0.2,recurrent_dropout=0.2,return_sequences=True)))
     #model.add(Bidirectional(LSTM(2,dropout=0.2,recurrent_dropout=0.2,return_sequences=True)))
-    model.add(Dense(32, activation='tanh'))
+    model.add(Dense(32, activation='relu'))
+    model.add(Dense(32, activation='relu'))
     model.add(Dropout(0.5))
     model.add(Dropout(0.5))
     model.add(Dense(len(CLASSES), activation='softmax'))
     model.add(Dense(len(CLASSES), activation='softmax'))
-    model.add(Activation('sigmoid'))
     
     
     # summarize the model
     # summarize the model
     print(model.summary())
     print(model.summary())
@@ -154,7 +154,7 @@ def classification_embedings_rnn(tweets_train, tweets_train_labels_numeric, twee
     # compile the model
     # compile the model
     model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['acc'])
     model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['acc'])
     print("Training the model...")
     print("Training the model...")
-    model.fit(train_features_pad, tweets_train_labels_numeric, batch_size=32, epochs=10, verbose=1, validation_data=(train_features_pad,tweets_train_labels_numeric))
+    model.fit(train_features_pad, tweets_train_labels_numeric, batch_size=32, epochs=50, verbose=1, validation_data=(train_features_pad,tweets_train_labels_numeric))
     loss, accuracy = model.evaluate(train_features_pad, tweets_train_labels_numeric, batch_size=32, verbose=1)
     loss, accuracy = model.evaluate(train_features_pad, tweets_train_labels_numeric, batch_size=32, verbose=1)
     print('Accuracy trainning: %f' % (accuracy*100))
     print('Accuracy trainning: %f' % (accuracy*100))
     
     

results_evaluation

+Experiment 1:
+
 Params: 
 Params: 
 
 
 vocab_embeddings = 100000
 vocab_embeddings = 100000
@@ -24,3 +26,30 @@ Macro-Recall: 0.5156404166549265
 Macro-F1: 0.5100454275009482
 Macro-F1: 0.5100454275009482
 Accuracy: 0.5155875299760192
 Accuracy: 0.5155875299760192
 
 
+
+Experiment 2: 
+
+Params: 
+
+vocab_embeddings = 100000
+max_lenght_tweet = 40
+ 
+Layers:
+
+e = Embedding(feature_size, EMBEDDING_DIM, input_length=max_len_input, weights=[embedding_matrix], trainable=False)
+model.add(e)
+model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2))
+model.add(Dense(32, activation='relu'))
+model.add(Dense(32, activation='relu'))
+model.add(Dropout(0.5))
+model.add(Dense(len(CLASSES), activation='softmax'))
+
+Accuracy trainning: 69.861719
+
+Results evaluation:
+
+*** Results RNN_LSTM ***
+Macro-Precision: 0.6040684610078678
+Macro-Recall: 0.6040370418815006
+Macro-F1: 0.6024329096077434
+Accuracy: 0.604003753518924