Deal with corpus data including reading the hotel comments data, the words dicts for naive_SA ...
Use dictionaries to predict the sentiment(1, 0, -1) of the given text.
Run:
python ./naive_SA.py
Use naive bayes to classify the text.
Run:
python ./bayes.py
Use CNN to do text classification.
Run textcnn.py to do train(train_data_path) first, the model will be saved in ./textcnn/model/
Then run textcnn.py to do test(test_data_path, r'model/xxx/checkpoints').