USE OF NAIVE BAYES ALGORITHM FOR SENTIMENT ANALYSIS ON THREADS APP REVIEWS ON GOOGLE PLAY STORE

Abstract

Advances in information technology encourage users to review digital applications, including Threads, a text-based platform released by Meta Platforms Using the Naive Bayes Algorithm, this study aims to analyze user reviews of Threads applications on the Google Play Store. The research stages include data collection through scraping techniques, data initial processing or pre-processing, weighting using TF-IDF techniques, and sentiment classification by dividing data into training and testing sets. The results of the analysis show that many of the discussions that use have positive sentiments, with a classification accuracy rate of 84%. For negative sentiment, precision, recall, and f1-score reached 74%, 65%, and 69%, respectively. Meanwhile, positive sentiment has a precision of 84%, a recall of 89%, and an f1-score of 86%. Although the model managed to classify sentiment well, there were still misidentification in negative reviews, caused by factors such as ambiguous language and mixed sentiment. This research aims to provide good information for developers to improve the quality of Threads applications in the user experience. For further research, it is recommended to compare other algorithms such as k-Nearest Neighbour or RBF kernel and apply the k-fold cross validation method to improve the accuracy of the model.