Unraveling Emotions: Contemporary Approaches in Sentiment Analysis

Abstract

Eberechukwu Q. Chinedu, Emmanuel Chibuogu Asogwa, Belonwu Tochukwu Sunday, Ndukaku Macdonald Onyeizu and Okechukwu J. Obulezi

Sentiment analysis, an essential component of natural language processing, plays a pivotal role in deciphering public opinion and emotional cues within the vast sea of user-generated content on social media platforms. This paper presents a focused analysis on sentiment analysis leveraging two different techniques, namely VADER (Bag of Words approach) and the RoBERTa model, an extension of the BERT architecture, known for its outstanding performance in a wide range of Natural Language Processing (NLP) tasks. Model A achieved a 47% accuracy in sentiment classification, while Model B demonstrated a higher accuracy of 73%. The findings not only highlight the performance disparities between the two models but also offer insights into the factors contributing to their varying degrees of accuracy. This research underscores the significance of model selection in sentiment analysis tasks and contributes to a better understanding of their applicability in real-world scenarios.

PDF