IDENTIFIKASI PERSPEKTIF PENGGUNA MENGGUNAKAN SENTIMENT ANALYSIS PADA APLIKASI MYSMARTFREN BERDASARKAN ALGORITME NAÏVE BAYES

Era of technology is increasingly developing, one of them is Smartfren Telecom.Tbk. Issued an innovation MySmartfren application, but the application is still considered less than optimal. Where each application allows users to provide reviews to the application. Sentiment analysis is a field of science that analyzes and processes opinions, emotions, and assessments by the public. The purpose of this study is to determine the accuracy of the data, find out the user's perspective and know the word patterns that often appear in MySmartfren reviews. The steps taken before the classification process are preprocessing or cleaning up the data, with the sequence of casefolding, tokenizing, filtering and stemming, then data is labeled positive and negative sentiments. Word cloud results get positive sentiment includes good, application, smartfren, love, receive, ok, easy, quota, signal, like. While negative sentiments include application, smartfren, signal, error, help, internet, quota, buy, network, 4G. In the next stage the naïve bayes classification method is used with the aim of obtaining data accuracy. The results obtained at a ratio of 50:50 is 79.96%, 60:40 is 85.66%, 70:30 is 88.8%, and the 80:20 ratio gets 93.83%. In addition, sentiment results obtained 1240 negative sentiments, and 240 positive sentiments while the result of MySmartfren user confidence tends to be negative, it is found that the words that are thrown with an average value of 0.867 compared to positive results are only 0.133 of the maximum total value of 1.