Application of the K-Nearest Neighbor (Knn) Algorithm in Sentiment Analysis of the Ovo E-Wallet Application

Abstract

Abstract— The OVO application can be downloaded on the Android platform via Google Play, Google play has a review feature on the application product to be downloaded, so that the review can be viewed or accessed by anyone, With these reviews, potential users of the application will see how important it is to consider using an application, problems regarding reviews or sentiment analysis of applications processed using text mining. The purpose of this study is to provide information to prospective OVO application users before using the application which can be seen from the results of giving reviews based on rating or stars (*) in the OVO application review column on Google Play and the authors categorize them into 3 classes, the first class ( 1 to 5 stars, second class (1 and 5 stars) third class by providing labeling grouping (1&2 stars are negative labels, 3 stars are neutral labels and 4&5 stars are positive labels) testing using the k-nearest neighbor method by finding the value of k from the k value of 1-10 to get the highest accuracy value, in order to obtain the highest accuracy value of 84.86% in the 2nd class test and giving a value of k 1 which means that the 1st and 5th star tests get positive values so that they can give a good impression to prospective application users OVO

Sentiment and Sunsets: Analysing Online Reviews of Kuta Beach in Bali

“…In addition, the study underscores the significance of Google Play reviews in shaping users' choices when it comes to using apps. It sheds light on the wider implications of Google reviews on consumer behaviour and decision-making (Masturoh et al, 2023). Research has shown that Google reviews have a substantial impact on tourists' choices and their overall impressions of tourist destinations (Fu & Pan, 2022).…”

Analisis Sentimen Pada Ulasan Access by Kereta Api Indonesia Dengan K-Nearest Neighbor

“…Dari pelabelan tersebut maka dapat didimpulkan positif sebanyak 555 data dan komentar negatif sebanyak 3160 data. Labeling rating pernah dilakukan oleh [12]…”