Analisa Sentimen Ulasan Aplikasi Wetv Untuk Peningkatan Layanan Menggunakan Metode Naïve Bayes

Abstract

The most popular online streaming application is WeTV. WeTV is an internet-based streaming service that is used by the public as an entertainment medium. The WeTV application has been downloaded by up to 50,000 users. Application user ratings may affect the image of the application depending on the services provided by the application developer. Many positive, neutral and negative reactions have had a big impact on WeTV. Categorizing user ratings cannot be done manually because it is not easy with very large amounts of data. Therefore, the purpose of this research is to analyze the user rating of the WeTV application on the Goggle Playstore. In this study the processing steps consisted of cleaning, case convolution, tokenization, normalization, stopword and vape removal, after which it was continued with the TF-IDF step and the final result was a confused matrix using the Python programming language with Naive Bayes classifier. method in this research. Using 12,000 reviews found on Google Playstore. to generate positive, negative and neutral sentiments from Wetv application user comments in the play store. The test with the highest precision value of 0.64% with a -1 precision value of 0.58% in Class Recall gives a value of 0.89% in the 90%:10 balance model.