A Survey on Twitter Sentiment Analysis Using Machine Learning Techniques

Twitter sentiment analysis involves various stages of the analysis process includes Text preprocessing techniques are applied to prepare the data, followed by examining the distribution of sentiment analysis. By utilizing the TF-IDF vectorizer, the textual data is transformed into numerical feature vectors. Three machine learning models, namely Bernoulli Naive Bayes (BernoulliNB), Linear Support Vector Classification (LinearSVC), and Logistic Regression, are created and evaluated using standard performance metrics like accuracy, precision, recall, and F1 score. The evaluation results effectively showcase the performance of each sentiment analysis model. The data is sourced from Twitter. The Logistic Regression model stands out in accurately classifying sentiments, while the LinearSVC and BernoulliNB models also exhibit high performance. The trained models are saved for future utilization, facilitating their integration into practical applications. This study presents a comprehensive approach to Twitter sentiment analysis, encompassing data preprocessing, model development, model evaluation, and storage for Twitter sentiment analysis tasks.