Aspect based Sentiment Analysis using RNN-LSTM

Abstract

Classifying user reviews into a certain category of interest such as positive or negative is known as Sentiment Analysis. User reviews can be their feedback about a particular product or service. In our work, feature based sentiment analysis is performed on film reviews.First these reviews are preprocessed and then polarity is determined. We have performed sentiment classification using the deep learning approach, LSTM (Long Short Term Memory) in Recurrent Neural Network (RNN) and compared it with machine learning approaches like Naive Bayes and logistic regression. Experimental results indicate that our model gives higher accuracy compared to other models.

Keywords: Sentiment Analysis, Aspect Based sentiment analysis, LSTM, RNN, Naive Bayes, Logistic Regression.