Contextual-cognitive convolutional attentive LSTM approach for sentiment analysis of Tamil review

Abstract

Movie goers use Twitter and other websites to share their thoughts and feelings regarding Tamil films in this age of social networking and online interaction. It is essential to examine these feelings for realizing the responses of the public and assess a film's level of success. In this research, a novel contextual-cognitive convolutional attentive long/short-term memory (C3-ALSTM) strategy for assessing Tamil movie reviews via Tamil tweets is presented. We first gather the Tamil movie review dataset through an open database to assess the suggested C3-ALSTM approach. We normalize the gathered data by eliminating extraneous information, noise and redundant data. Tokenization and the removal of stop words and special characters are employed in this phase. Following data cleansing, term frequency-inverse document frequency (TF-IDF) is applied for extracting the attributes. The suggested approach evaluates the Tamil reviews based on these retrieved features. The suggested approach is put into practice using the Python platform and its accuracy, precision, recall and f-measure metrics are examined. According to the research's findings, the suggested method outperforms other used techniques for sentiment assessment of Tamil reviews. In addition, this research attempts to offer insightful information about how the public reacts to particular Tamil films, advancing knowledge of public preferences and attitudes toward the Tamil cinema industry.