Application of the Naïve Bayes Algorithm in Sentiment Analysis of Using the Shopee Application on the Play Store

Abstract

This research aims to find out the opinions of users of the Shopee application on the Play Store using the Naive Bayes Naive Algorithm and to find out the suitability of the correct application of the Naive Bayes algorithm in carrying out sentiment analysis with the classification of three sentiment classes. The dataset used in this study consisted of 2000 customer reviews obtained from the Play Store in 2024 collected by the scraping process using the Python library. The dataset has 1,198 examples of negative attitudes, 583 examples of good sentiment, and 219 examples of neutral sentiment. The results of this study are expected to be used as evaluation material for Shopee Apilkation to improve the performance of Shopee applications. Research findings show that the Bayes naive approach reaches accuracy determined by various aspects, such as the quantity of data collections and positive and negative data distribution. This study shows that the Bayes naive algorithm can function properly as a technique to evaluate user sentiment for applications in the Play Store. However, with the classification of three classes, another algorithm is needed to produce higher accuracy.

Keywords

• sentiment analysis
• naive bayes algorithm
• shopee
• play store
• user reviews