Deciphering Social Media Sentiment for Enhanced Analytical Accuracy: Leveraging Random Forest, KNN, and Naive Bayes

The research aims to determine the most effective algorithm for social media sentiment analysis, evaluating Random Forest, K-Nearest Neighbor (KNN), and Naive Bayes models. This investigation holds importance as it seeks to unveil the key factors influencing sentiment analysis on social media using rigorous methodologies. Employing a dataset obtained from Kaggle, the study examined the predictive capabilities of Random Forest, KNN, and Naive Bayes in accurately predicting social media sentiment. Notably, Naive Bayes emerged as the most precise predictor of social media sentiment compared to Random Forest and KNN algorithms. The study's results highlight the potential of machine learning algorithms in providing accurate predictions for social media sentiment analysis. A statistically significant level of p= 0.04, where $\mathrm{p}\lt0.05$, further underscores the significance of the study, indicating notable variations among the three algorithms. With mean accuracies of 86.06% for Random Forest, 89.28% for KNN, and 92.01% for Naive Bayes, the findings affirm Naive Bayes' superior performance over the Random Forest-KNN combination, emphasizing its superior predictive accuracy.