The Effectiveness of the Ensemble Naive Bayes in Analyzing Review Sentiment of the Lazada Application on Google Play

The surge in e-commerce growth in Indonesia has led to the emergence of numerous new online marketplaces, including the Lazada application. Within the application, users encounter various experiences and can share their insights through reviews, highlighting both strengths and weaknesses. However, the abundance of user reviews makes it challenging to extract pertinent information tailored to individual needs. Consequently, employing sentiment analysis becomes a viable solution to sift through user review data, providing a thorough assessment of the Lazada application's quality. The methods used in this research are Term Frequency - Inverse Document Frequency (TF-IDF) and Ensemble Learning, specifically utilizing a Voting approach. The research is used to compare the effectiveness between single Naïve Bayes and the Voting method. Multinomial Naïve Bayes demonstrates higher accuracy and F1Score compared to other classification models, achieving 89.1% accuracy and 89.65% F1-Score.