Analisis Sentimen Ulasan Pengguna Aplikasi Grab Mobile Menggunakan Metode K-Nearest Neighbor dan Lexicon-Based

Abstract

Grab is one of the most widely used online transportation and digital service applications in Indonesia. As the number of users grows, reviews provided on the Google Play Store have become an important sourc of information to understand what users think and how satisfied they are. This study aims to study customer feelings about app reviews. Grab uses the K-Nearest Neighbor (KNN) method combined with a Lexicon -Based approach for automatic labeling. The dataset consists of 300 reviews in Indonesian sourced from the Google Play Store. The preprocessing process includes cleaning, case folding, tokenizing, stopword removal, and stemming using the Sastrawi library. Emotion labeling is done automatically using a Lexicon-Based sentiment dictionary. Text attributes are extracted using the TF-IDF (Term Frequency-Inverse Document Frequency) method, then classified using KNN  with K = 5 and a training and experimental data sharing ratio of 80: 20. The research findings emphasize that the majority of users (80.67%) give positive reviews to the Grab app. The KNN model achieved 90% accuracy with a precision of 0.92, a recall of 0.96, and an F1-score of 0.94 for the good class. This study demonstrates that the combination of KNN and Lexicon-Based methods can be used effectively in sentiment classification of Indonesian-language reviews.