Sentiment Analysis of OYO App Reviews Using the Support Vector Machine Algorithm

Abstract

The rapidly growing tourism industry causes the need for hotels to increase. This has led to innovation in the form of virtual hotel operators, one of which is OYO. OYO is a VHO application with the highest rating found on the Google playstore. OYO is one of the applications with millions of users. Of course, this cannot be separated from the ratings and reviews from users. The reviews contained in the playstore itself can contain positive, neutral or even negative opinions. Given the importance of user reviews to application development, this research classifies reviews on the OYO application to determine user sentiment. In this study, the data used is 2,000 data which will be classified into positive, neutral and negative sentiments. The Support vector machine algorithm was chosen because it is capable of producing high accuracy. Based on testing, the Radial basis function kernel is able to produce the highest accuracy among other kernels and by using a dataset division ratio of 80:20 the accuracy obtained is 78.98%. While testing using the Confusion matrix produces an accuracy of 80.36%.