Android Application For Analysis Review On Google Playstore Using Support Vector Machine Method
Andrean Setiawan, Viny Christanti Mawardi · 2022 5th International Conference on Information and Communications Technology (ICOIACT) · 2022
Nowadays, Google Play Store have more than 2.77 million of mobile applications. Though there are several other app stores available for Android, none of them compare to the Google Play Store. Many people who use the application would leave their review after using the application at Google Play Store. Google Play Store also be used to do research from their review. Information that related to advantages, features and other functions may not necessarily be explained in detail by the application developer. Reviews from the users can be used as illustration of the experience from using an application for potential users who will download the application. This study aims to provide information to users regarding the results of sentiment analysis of user reviews from applications on the Google Play Store and sort applications in certain categories based on the best results of the sentiment analysis. This study used 500 data review from each application based on the most relevant filter. The classification results will be evaluated using a confusion matrix. The analysis results are displayed in an android application that build using the JavaScript programming language with react native library. From the test results, the best kernel for performing calculations is the RBF kernel with a combination of 80% training data and 20% test data. The accuracy value is 73.97%, the precision value is 76.53%, the recall value is 71.57% and the F1 score is 73.35%.