Performance Comparison of Supporting Vector Machine Method without or with Particle Swarm Optimization Based on Sentiment Analysis WhatsApp Review

Communication technology is undergoing a very rapid development. With this development, it allows humans to communicate remotely using applications available on smartphones. WhatsApp as one of the applications for communicating remotely that offers various advantages to facilitate human communication through smartphones. In addition to its advantages, WhatsApp also has various disadvantages that can be seen in the reviews on the Google Play Store. Reviews on the Google Play Store can be an illustration of the app's eligibility for new users to download. Reviews of the WhatsApp application can be done using sentiment analysis. This study aims to classify positive and negative sentiments on WhatsApp application reviews and compare the performance of the SVM algorithm without and based on PSO. The method used in this study is the method of supporting vector machines without and based on PSO. In the results of the weighting of words using TF-IDF obtained the word "original" with a weight of 0.32. Then the test results, the accuracy of the Support Vector Machine was 79% and the Support Vector Machine based on PSO was 80%. Obtained 1431 positive sentiments and 1069 negative sentiments on the review of the WhatsApp application on the Google Play Store. Based on the results of the paired t-test, it was obtained that there was a significant difference in values between the SVM accuracy values without and based on PSO. The performance of PSO-based SVM algorithms has higher accuracy than the performance of SVM algorithms without PSO.