Attention LSTM framework for Aspect-enabled Sentiment Analysis

Abstract

Sentiment analysis is the automated analysis of the input data and identifies the person's attitude from the text. Aspect-enabled Sentiment Analysis demonstrates the relationship between the opinion aims of a particular document and the polarity values between the texts while aspects are implied, it is a very complex job to analyse and compute the corresponding polarity. In modern days, various techniques and enhancements have been involved in efficiently solving these kinds of issues. At the same time, the correlated aspects through the pre-defined groups could scuffle when the involvement of the low-performed aspects. The Attention LSTM framework is incorporated in this proposed work for performing the sentiment analysis the attention LSTM framework could focus on various sentence parts while the different kinds of aspects are involved in the input text and the IMDB movie review dataset is used for aspect-enabled sentiment analysis. Experimental results show that the proposed technique has produced 94.5% accuracy and the relevant techniques of SVM, RNN, and LSTM in various performance metrics.