A Voting-Based Sentiment Classification Model

Abstract

Sentiment analysis is used to depict sentiments present in the text structures, including news, reviews, and articles, and classify them as positive, or negative. It has gained significant attention due to the increase in individuals utilizing social media platforms to express sentiments about organizations, products, and administrations. Many methods are being devised to improve the efficacy of automated sentiment classification. The study proposes a voting-based ensemble model Majority Voting (MV) using five supervised machine learning classifiers named Logistic Regression (LR), Support Vector Machine (SVM), Artificial Neural Network (ANN), Decision Tree (DT), and Random Forest (RF) as base classifiers and a majority voting rule-based mechanism to get the final prediction. The performance of the proposed method is assessed using minimum, maximum, mean, and median values of precision, recall, f-score, and accuracy. The results of 900 values of the classification accuracy (3 datasets * 6 (classification methods) * 10 data subsets (k-fold cross-validation for $k=10$) * 5 runs), indicates that the proposed approach outperforms the individual classifiers in majority of the cases.