Ensemble-based method designed to Increase Sentiment Analysis Categorization Accuracy

Abstract

The need for understanding user behaviour is strong due to the ever-increasing volume of data generated by social media users, especially in light of the current coronavirus outbreak. In this experiment, we focus on a dataset that includes the subjective assessments of those who have written about the epidemic. It is not easy to find the best classification methods for this sort of data. When compared to traditional feature-based methods, deep learning models for sentiment analysis may provide more nuanced representations and better overall performance. The suggested method is an ensemble strategy that makes use of boosting in addition to supervised machine learning methodologies. The goal is to examine the usefulness of the idea of using several classifier systems on IMDb and other multi-domain datasets. We have seen implementations of the Vote method in combination with Naive Bayes, Maximum Entropy, and Boosting classifiers. The Ensemble method outperforms both the best-reported individual classifier (Support vector machines) and Naive Bayes (which is widely used). Precision, recall, and accuracy are only few of the metrics that are used to evaluate the effectiveness of various approaches.