Sentiment Analysis Using Machine Learning and Deep Learning Models

This study investigates sentiment analysis method-ologies using two distinct datasets: the IMDb Movie Reviews corpus and a novel Financial Sentiment dataset. Employing a comprehensive approach, the research integrates traditional Machine Learning algorithms and Deep Learning models to analyze and interpret sentiments. The preprocessing phase incor-porates techniques such as punctuation removal, text lowercasing, and specialized considerations for financial symbols. Machine Learning models, including Logistic Regression and Random Forest, are applied to the IMDb dataset, with Naive Bayes demonstrating superior performance at 89% accuracy. The Deep Learning model achieves 82% accuracy on the testing set. In the Financial Sentiment dataset, Naive Bayes emerges as the optimal model, despite reduced overall performance due to dataset limitations and imbalance. This research contributes valuable insights into sentiment analysis methodologies, offering a nuanced understanding of model performance across diverse datasets and enhancing our comprehension of sentiment classification in various contexts.