IMDB-text-classification-sentiment-prediction
Easy to start.
Use deep nerual network to predict the sentiment of movie review.
Various methods, word2vec, tf-idf and df to generate text vectors.
Various models including lstm and cov1d.
Achieve f1 score 92.
- Run python text_wash.py in the data_process folder
- Run python data_process_word2vec.py, python data_process_tf_manual.py, python data_process_tfidf.py under the data_process folder, this step takes a long time
- Use bash main.sh word2vec to perform text classification training using word2vec word vectors as training data, and use bash view.sh word2vec to view the redirected output during training
- Use bash eval.sh word2vec to test the saved model after training
The above word2vec can be replaced by (word2vec/tf/tfidf)