Performance analysis of Sentiment Classification using Optimized Kernel Extreme Learning Machine
D. N. S. B. Kavitha, M. Venkata Subbarao · 2023
Experts’ attention has recently shifted to sentiment classification because of its applications in a variety of fields. In the past two decades, scientists have developed various multidisciplinary methodologies for sentiment analysis (SA) to enhance classification accuracy. Initial studies in sentiment analysis concentrated mostly on structured customer product and service reviews. However, academics have abruptly shifted their focus to unstructured data because of the availability of real-time messages via Twitter. Information considered to be more valuable is taken from tweets that express a point of view. There are a variety of programs, such as sentimentor, designed expressly for assessing the feelings of tweets. This paper provides a comprehensive overview of the several available SA methodologies, from the earliest to the most recent and accurate. Further the paper presents emotion detection in political tweets about an individual using various supervised machine learning (ML) approaches under different training rates. Kernel Extreme Learning Machine (KELM) by incorporating it with the Salp swarm optimization algorithm (SSA) is investigated under different training rates and with different data sets. Experimental results depicts that proposed KELM optimized by SSA attains improved performance compared to KNN, SVM, and Ensemble Classifiers.