ran unigrams with lr 0.5 and nest 100 weighted

AIclassifier.py

@@ -57,7 +57,7 @@ def format_sentVector_to_SparseMatrix(vectors_list, dictionary):
         counts = Counter(sent_vector)
         for index, freq in counts.items():
             if len(counts.items()) > 0:
-                sparse_vector[index] = 1 #freq/len(sent_vector) # 1 DIFFERENT CONFIGURATION POSSIBILITIES
+                sparse_vector[index] = freq/len(sent_vector) # 1 DIFFERENT CONFIGURATION POSSIBILITIES
         if (i == 0): # TO DO: OPTIMIZE, NO NEED TO CHECK THIS EVERY TURN
             matrix_array = [sparse_vector]
         else:
@@ -119,11 +119,11 @@ def create_vectors_list(sents, conversion_dict):
                 sent_mixed_vector = sent_mixed_vector.astype(int)
         vectors_list.append(sent_vector)
         bigrams_vector.append(sent_bigrams_vector)
-        print(sent_mixed_vector)
+        #print(sent_mixed_vector)
         mixed_vector.append(sent_mixed_vector)
 
-    return vectors_list # unigrams
-    #return bigrams_vector
+    #return vectors_list # unigrams
+    return bigrams_vector
     #return mixed_vector
 
 # def create_word_embedding(partition):
@@ -197,8 +197,8 @@ def create_vectors_list(sents, conversion_dict):
 #train_word_embedding_features = create_word_embedding(sents_train)
 #dev_word_embedding_features = create_word_embedding(sents_dev)
 #test_word_embedding_features = numpy.asarray(create_word_embedding(sents_test))
-print("Length of the dictionary of word representations:",len(words_to_numbers))
-#print("Length of the dictionary of word representations:",len(bigrams_to_numbers))
+#print("Length of the dictionary of word representations:",len(words_to_numbers))
+print("Length of the dictionary of word representations:",len(bigrams_to_numbers))
 #print("Length of the dictionary of word representations:",len(mixed_to_numbers))
 
 # FLAG - CHECK IF DICTIONARY IS BUILT CORRECTLY
@@ -207,13 +207,13 @@ def create_vectors_list(sents, conversion_dict):
 # count frequency before and after removing unknown words - ??? - ASK GABRIEL!!
 # checked that it seems ok
 
-train_vectors_list = create_vectors_list(sents_train, words_to_numbers)
-dev_vectors_list = create_vectors_list(sents_dev, words_to_numbers)
-test_vectors_list = create_vectors_list(sents_test, words_to_numbers)
+# train_vectors_list = create_vectors_list(sents_train, words_to_numbers)
+# dev_vectors_list = create_vectors_list(sents_dev, words_to_numbers)
+# test_vectors_list = create_vectors_list(sents_test, words_to_numbers)
 
-# train_vectors_list = create_vectors_list(sents_train, bigrams_to_numbers)
-# dev_vectors_list = create_vectors_list(sents_dev, bigrams_to_numbers)
-# test_vectors_list = create_vectors_list(sents_test, bigrams_to_numbers)
+train_vectors_list = create_vectors_list(sents_train, bigrams_to_numbers)
+dev_vectors_list = create_vectors_list(sents_dev, bigrams_to_numbers)
+test_vectors_list = create_vectors_list(sents_test, bigrams_to_numbers)
 
 # train_vectors_list = create_vectors_list(sents_train, mixed_to_numbers)
 # dev_vectors_list = create_vectors_list(sents_dev, mixed_to_numbers)
@@ -223,13 +223,13 @@ def create_vectors_list(sents, conversion_dict):
 
 # FLAG - CHECK IF SENTENCE REPRESENTATIONS WERE DONE CORRECTLY
 
-train_matrix_array = format_sentVector_to_SparseMatrix(train_vectors_list, words_to_numbers)
-dev_matrix_array = format_sentVector_to_SparseMatrix(dev_vectors_list, words_to_numbers)
-test_matrix_array = format_sentVector_to_SparseMatrix(test_vectors_list, words_to_numbers)
+# train_matrix_array = format_sentVector_to_SparseMatrix(train_vectors_list, words_to_numbers)
+# dev_matrix_array = format_sentVector_to_SparseMatrix(dev_vectors_list, words_to_numbers)
+# test_matrix_array = format_sentVector_to_SparseMatrix(test_vectors_list, words_to_numbers)
 
-# train_matrix_array = format_sentVector_to_SparseMatrix(train_vectors_list, bigrams_to_numbers)
-# dev_matrix_array = format_sentVector_to_SparseMatrix(dev_vectors_list, bigrams_to_numbers)
-# test_matrix_array = format_sentVector_to_SparseMatrix(test_vectors_list, bigrams_to_numbers)
+train_matrix_array = format_sentVector_to_SparseMatrix(train_vectors_list, bigrams_to_numbers)
+dev_matrix_array = format_sentVector_to_SparseMatrix(dev_vectors_list, bigrams_to_numbers)
+test_matrix_array = format_sentVector_to_SparseMatrix(test_vectors_list, bigrams_to_numbers)
 
 # train_matrix_array = format_sentVector_to_SparseMatrix(train_vectors_list, mixed_to_numbers)
 # dev_matrix_array = format_sentVector_to_SparseMatrix(dev_vectors_list, mixed_to_numbers)
@@ -285,9 +285,9 @@ def create_vectors_list(sents, conversion_dict):
 importances = model.feature_importances_
 
 features = {}
-for i,(token,value) in enumerate(zip(words_to_numbers, importances)):
+#for i,(token,value) in enumerate(zip(words_to_numbers, importances)):
 #for i,(token,value) in enumerate(zip(mixed_to_numbers, importances)):
-#for i,(token,value) in enumerate(zip(bigrams_to_numbers, importances)): # IMPORTANTO TO CHANGE TO ADEQUATE DICT
+for i,(token,value) in enumerate(zip(bigrams_to_numbers, importances)): # IMPORTANTO TO CHANGE TO ADEQUATE DICT
    if (value != 0):
        features[token] = value
 features = sorted([(value, key) for (key, value) in features.items()], reverse=True)
@@ -326,7 +326,7 @@ def create_vectors_list(sents, conversion_dict):
 os.makedirs(os.path.dirname(path), exist_ok=True)
 with open(path, 'w') as file:
     #print("Performance measures - Unigram Dictionary - MLP Word Embeddings\n", file=file)
-    print("Performance measures - Unigram Dictionary - Adaboost\n", file=file)
+    print("Performance measures - Bigram Dictionary - Adaboost\n", file=file)
 #write_output_stats_file(path, "Dev", dev_labels_primary, predictions, labels)
 write_output_stats_file(path, "Test", test_labels_primary, predictions, labels)
 

output/AI Classifier/1labelPredictionsStatsTest.txt

@@ -2,13 +2,13 @@ Performance measures - Unigram Dictionary - Adaboost
 
 Test set:
 
-Precision macro: 0.247
-Precision Individually: [0.333 0.    0.566 0.    0.333]
-Recall macro: 0.256
-Recall Individually: [0.214 0.    0.968 0.    0.1  ]
-F1 Score micro: 0.507
-F1 Score macro: 0.226
-F1 Score weighted: 0.408
-F1 Score Individually: [0.261 0.    0.714 0.    0.154]
+Precision macro: 0.346
+Precision Individually: [0.5   0.    0.564 0.    0.667]
+Recall macro: 0.283
+Recall Individually: [0.214 0.    1.    0.    0.2  ]
+F1 Score micro: 0.537
+F1 Score macro: 0.266
+F1 Score weighted: 0.442
+F1 Score Individually: [0.3   0.    0.721 0.    0.308]