Comparative Analysis of Sentiment Analysis Models on Twitter Data Using Machine Learning

Sentiment analysis is a machine learning and Natural Language Processing (NLP) technique that classifies the text into positive, negative and neutral emotions based on tone of the text. This research uses different sentiment analysis models such as Roberta-an enhanced BERT Model, VADER and Logistic Regression to analyze the sentiment of user tweets within Twitter dataset that is publicly available in Kaggle. The dataset consists of a large number of tweets, including emojis and retweets. To avoid any bias and inconsistencies, the data is processed by handling retweets and emoji’s. For all the models the same preprocessing technique is applied on the data making sure that each model performance is accurately measured. Each of the models are then applied to the data to classify the sentiment of the tweet. The results obtained from each model are then compared, which highlights how well each model performs in relation to sentiment analysis. These findings offer valuable information on the benefits and drawbacks of each technique, and they also provide important recommendations for future application in related domains. Additionally, this suggested study is beneficial for researchers who are interested in doing sentiment analysis on social media data.