Enhanced sentiment analysis and emotion detection in movie reviews using support vector machine algorithm

Abstract

Films evoke diverse responses and reactions from audiences, captured through their reviews. These reviews serve as platforms for audiences to express opinions, evaluations, and emotions about films, reflecting the personal experiences and unique perceptions of the viewers. Given the vast volume of reviews and the distinctiveness of each perspective, automated analysis is essential for efficiently extracting valuable insights. This study employs the support vector machine (SVM) algorithm for classifying movie reviews into positive and negative categories. The dataset includes 50,000 IMDb movie reviews, split evenly between positive and negative sentiments. Each review is analyzed using the National Research Council Canada (NRC) emotion lexicon (NRCLex) to assign scores for emotions such as anger, disgust, fear, joy, sadness, and surprise. Subsequently, these reviews are further analyzed using term frequency-inverse document frequency (TF-IDF) for classification. The proposed algorithm achieves 90% accuracy, indicating its effectiveness in classifying sentiments in movie reviews. The study's findings confirm the potential of the SVM algorithm for broader applications in sentiment analysis and natural language processing. Additionally, integrating emotion detection enhances understanding of nuanced emotional content, providing a comprehensive approach to sentiment classification in large datasets.