Swarm intelligence based deep learning and ensemble multi models for product review sentiment analysis

Abstract

The growth of user-generated content in websites and social networks, newlinee-commerce such as Amazon, and Trip Advisor, has led to an increasing use newlineof social networks for expressing opinions about services, products or events. newlineSentiment analysis is used to extract the features or aspects of the user by newlineanalyzing and classifying the text posted by social media and websites. newlineAspect-Based Sentiment Analysis (ABSA) system is the best solution for newlineefficient analysis of user reviews. The ABSA system identifies the sentiments newlinefor each attribute at a fine granular level, which assists the decision process newlinefurther effectively than previous SA models. In this aspect extraction is the newlinemain process that classifies the user aspects. Earlier, Neural Network models newlinewere employed in ABSA but in complex comments the word features which newlinemight lead to loss of key text information. It often ignores context newlineinformation and the semantics of words, which degrade the accuracy of newlinesentiment analysis. Due to the powerful feature extraction ability, deep neural newlinenetwork bring new potential for sentiment analysis, which can better learn newlinecontext information and the semantics of words. Deep learning methods have newlinebeen applied in the field of product reviews to achieve satisfactory accuracy. newlineThus, designing an effective method for product review sentiment analysis newlinebecomes a major important task. newline