Benchmarking Machine Learning Methods for Sentiment Analysis in Social Media: A Comprehensive Investigation

Because so many people utilize the Internet, social media has ingrained itself into our daily lives. Globally, Facebook and Twitter are among the most widely used social media networks. The goal of sentiment analysis is to find user posts on social media on a certain subject and classify them as neutral, positive, or negative. Therefore, the purpose of the study is to look into how different text representation formats affect sentiment analysis performance. The procedures utilised in the present research employed two datasets. The first is made up of Facebook comments from users, and the second is made up of tweets from Twitter users. The research employed the Python programming language to create classification models that classified views into positive, negative, and neutral classifications using the Random Forest (RF), Decision Tree (DT), Super vector machine (SVM), and K-nearest Neighbours (KNN) algorithms. At the conclusion, the categorization algorithms’ success rates were contrasted. This study looks at the subject and suggests an algorithm-based sentiment analysis solution. The results of the experiment showed that Random Forest (94.3% for Facebook dataset and 98.3% Twitter dataset) fared better in terms of accuracy in both datasets than the other models.