ANALYSIS OF GOOGLE CAPCUT APP USER REVIEWS USING NAIVE BAYES

Abstract

In the current digital landscape, video editing platforms such as CapCut have emerged as vital tools for content creators across various social media channels. Feedback from users—whether positive, negative, or neutral—serves as a critical resource for guiding application improvement. This research seeks to examine user sentiment toward CapCut through the application of the Naive Bayes classification algorithm, recognized for its efficiency and straightforward implementation in processing textual data. User review data was obtained via web scraping from the Google Play Store, yielding a dataset of 100 entries. The study was conducted following the Knowledge Discovery in Databases (KDD) framework, which includes stages of data selection, preprocessing, text transformation, classification using Naive Bayes, and assessment through performance metrics such as accuracy, precision, recall, and F-measure. Results demonstrate that the algorithm successfully distinguishes between sentiment categories, while sentiment visualization through word clouds highlights the most frequently occurring terms. The outcomes of this study offer practical insights for developers seeking to enhance application quality, as well as scholarly contributions to the fields of sentiment analysis and machine learning–based text mining.