SENTIMENT ANALYSIS AGAINST THE DANA E-WALLET ON GOOGLE PLAY REVIEWS USING THE K-NEAREST NEIGHBOR ALGORITHM

Abstract

DANA e-Wallet or digital wallet application can be downloaded on the Android platform via Google Play, and google play itself provides a review column. The public will usually see reviews on Google Play before they download an application because the information obtained through these reviews is considered effective in providing information, problems regarding reviews or sentiment analysis of the application must be processed using text mining. Text mining in this study uses k-nearest neighbor by testing 3 classes based on star rating, the first class consists of 1-5 stars, the second class consists of (1 & 5 stars, 3rd class consists of labeling stars (1 & 2) negative label, 3 neutral labels, as well as 4 & 5 stars positive labels) and testing the value of k 1-10 so that the highest accuracy value is obtained with class 2 (1 star and 5 stars) and the best test at k 1 value is obtained the accuracy result of 86.64%

Keywords

• e-wallet
• dana
• k-nearest neighbor