Classification of Google Play Application Using Decision Tree Algorithm on Sentiment Analysis of Text Reviews

Abstract

The study begins with a comprehensive background, examining the significance of classifying Google Play Store applications using user reviews. The problem statement revolves around the need for an efficient approach to classify application reviews to their sentiments, such as the user’s downloaded application not working as intended, and it can take time for users to read every single review. The project's objectives are to study the classifying approach of Google Play Store application reviews using the Decision Tree algorithm, develop a prototype of a classifying application program, and evaluate the accuracy model of the review classification program. To achieve these objectives, the methods employed involve data preprocessing and implementing the Decision Tree (DT) algorithm for classification. The classification model is trained and tested using various split ratios, and the optimal depth for the DT is determined through parameter tuning to achieve the best accuracy. Key results indicate that the developed prototype effectively classifies Google Play Store application reviews with an overall accuracy of 84.88%. This study has successfully achieved its objectives in creating a working Google Play application classification program. The classifier’s accuracy and user-friendly interface make it a valuable tool for developers and users.