Data Categorization and Review Identification on Twitter Using WordNet Implicit Aspect Sentiment Analysis

Abstract

Social media review analysis has developed into a fascinating profession that addresses important public safety issues that are respected globally. Sentiment analysis (SA) on Twitter is still a topic of ongoing attention in this profession. Tweet datasets for sentiment opposition bracket are subjected to aspect-grounded SA, a method that allows information to be extracted, dissected, and categorized in order to predict social media evaluations. The implicit aspect for social media review tweets that is implied by adjectives and verbs is the subject of this paper's aspect identification job. In order to improve training data for [1) Social media review Implicit Aspect Rulings Discovery (IASD) and Social media review Implicit Aspect Identification (IAI), a mongrel model is suggested. It is based on WordNet semantic relations and the Term-Weighting scheme. Three classifiers—Multinomial Naïve Bayes, Support Vector Machine, and Random Forest—are used to estimate the performance on three Twitter social media review datasets. The obtained results show the value of verbs in training data enhancement for social media review IASD and IAI, as well as the efficacy of WN reversal and description relations.