Comparative Analysis of Machine Learning Algorithms for Conducting Sentiment Analysis on Product Reviews Derived from Social Media Data

Sentiment analysis has garnered considerable attention in recent years, primarily on account of the exponential expansion of user-generated content on social media platforms and the escalating demand for businesses to extract valuable insights from this data. This manuscript presents a comprehensive comparative examination of different machine learning algorithms employed in the task of sentiment analysis on product evaluations obtained from social media data. To execute the comparative analysis, a diverse dataset of product evaluations from various social media platforms is assembled and processed. Subsequently, this dataset utilizes a range of varied machine learning algorithms. The effectiveness of each algorithm is assessed through the use of fundamental metrics including F1-score, precision, accuracy and recall. The findings of the investigation uncover the merits and deficiencies of each computational procedure when employed in sentiment analysis tasks involving data derived from social media. This scientific inquiry delivers invaluable perspectives into the selection of machine learning procedures for sentiment analysis on social media-based evaluations of products. It functions as a manual for practitioners and scholars in the process of choosing the most appropriate procedure based on the unique attributes of their data and the goals of their analysis. Furthermore, it contributes to the broader comprehension of sentiment analysis and its significance in utilizing social media data for making business decisions.