Sentiment Analysis Using Large Language Models: A Case Study of GPT-3.5
摘要
Sentiment analysis, which utilizes machine learning, natural language processing, and computational linguistics, has been developed to comprehend the emotions and viewpoints of individuals on social media platforms. This area of research has undergone extensive investigation, leading to the integration of diverse algorithms and techniques specifically tailored to this objective. A recent breakthrough in the field of general AI is exemplified by Large Language Models, which have surpassed numerous research domains as quintessential achievements. This study examines the performance of GPT-3.5 in sentiment analysis. We evaluate its performance on established benchmark datasets as well as a collection of reviews extracted from Google Maps. The findings indicate that the Large Language Models can outperform traditional methods in the literature, and they do not require pre-processing, unlike traditional methods.