def save_binary(words, filename, parsed_texts): # , predictions):
print("Saving to binary file.")
vec = CountVectorizer(analyzer="word", vocabulary=words)
train_data_features = vec.fit_transform(parsed_texts)
features_arr = train_data_features.toarray()
features_arr = np.sign(features_arr)
i = 0
for row in features_arr:
i += 1
if i > 200:
break
print(features_arr)
p_save(features_arr, filename)
#for i in range(NUM_FEATURES):
# print(' ' + cv.get_feature_names()[music[i]]);
#print('Interview: ');
#for i in range(NUM_FEATURES):
# print(' ' + cv.get_feature_names()[interview[i]]);
# Sum up the mutual information for all classes
miSum = mi.sum(0);
sortedIndices = np.argsort(-np.array(miSum[0])[0]);
features = [];
for i in range(NUM_FEATURES):
feature = cv.get_feature_names()[sortedIndices[i]];
# print(' ' + feature);
features.append(feature);
p_save(features, "mi_features.dat");
print("Loading dataset.")
data = []
ifile = open('data/ml_dataset_train.csv', "r")
reader = csv.reader(ifile)
i = 0
for row in reader:
if i == 0:
i = 1
continue
data.append(row)
ifile.close()