Application of XGBoost Algorithm in Sentiment Classification of MOBA Game Reviews on Google Play Store

In the rapidly evolving digital era, business applications like GoBiz play a crucial role in supporting the operations of Micro, Small, and Medium Enterprises (MSMEs). This study aims to analyze user sentiment toward the GoBiz app based on reviews on the Google Play Store by applying two machine learning algorithms: Extreme Gradient Boosting (XGBoost) and Random Forest. Two labeling approaches were used: score-based labeling, which refers to star ratings, and lexicon-based labeling using the VADER method. Data from 10,000 reviews were collected through web scraping and processed through preprocessing, labeling, TF-IDF feature extraction, model training, and evaluation. The evaluation results showed that the XGBoost algorithm excelled in score-based labeling with the highest accuracy of 86.81%, while Random Forest was more stable than the VADER approach with an accuracy of 84.98%. Both models performed well, but their effectiveness depended on the type of labeling used. This research contributes to the development of a sentiment classification system in digital business applications, and can be utilized by GoBiz application developers to improve service quality based on user perceptions.