Sentiment Analysis Based On User Reviews in the Gojek Application Using K-Nearest Neighbors (KNN)

Abstract

User reviews can provide valuable insights for business owners in responding to consumer feedback regarding their products. Sentiment analysis is crucial in making informed decisions to address user complaints, especially those found in Google Playstore reviews. By employing the K-Nearest Neighbors (KNN) algorithm to classify data containing negative sentiment, the IT team can better anticipate the complaints users experience while using the Gojek application. This study's findings indicate that the number of negative sentiments, totaling 223 reviews, surpasses the positive sentiments, which amount to 77 reviews in the amount of 300 reviews, with the highest accuracy 77% determined based on Confussion Matrix calculations. This information can then be utilized to design more effective and user-oriented update strategies, thereby enhancing the overall quality of the Gojek application.