This project contains several machine learning models trained from Large Movie Review Dataset.
This flask web api can predict the sentimental tendency (positive or negative) and highlight sentimental keywords based on your input text. Logistic regression (accuracy 0.877), support vector classifier (accuracy 0.865) and naive-bayes (accuracy: 0.84) are included. If you want to try out please download nlp_flask and type in the terminal python runserver.py
requirements: keras, nltk (download stopwords), sklearn, pickle, pandas
A LSTM RNN model is developed, please refer to the jupyter notebook.