WITHDRAWN: NA-DLSTM – A neural attention based model for context aware Aspect-based sentiment analysis

Abstract

Sentiment Analysis makes use of natural language processing approaches to methodically recognize, mine, and learn user sentiment. As such it can be useful for finding the mood of the public about a specific topic. Aspect based sentiment analysis is an important technique that analyses and breaks down text into aspects which are nothing but attributes or constituents of a product or service and then allocates a sentiment level (positive, negative or neutral) for each aspect. The difference between Sentiment analysis and Aspect based sentiment analysis is that the former identifies only the sentiment of an overall text, while the latter examines and categorizes the entire text into various aspects and finds the corresponding sentiment for each one. But sentiment context for a given aspect also plays an important rule for correct sentiment prediction. So, we developed a new model which is a combination of neural attention mechanism and LSTM called NA-DLSTM (Neural Attention based Deep Long Short-Term Memory) for Context-aware Aspect based Sentiment Analysis. We have compared our model with baseline models in terms of accuracy and F1 score on datasets (Rest14, Rest15, Rest16).