Analisis Sentimen Terhadap Sistem Informasi Akademik STIMIKOM Stella Maris Sumba Menggunakan Algoritma Naïve Bayes

Abstract

Stimikom stella maris sumba already has an academic information system, so that students can see the value without having to come to college, But there are some constraints that students often experience, among which there are often messages of terror when accessing the system and often error inclusion of the nim.According to the problem the researchers found it necessary to research sentimental analysis of student comments using naive bayes algorithms.The results that were obtained based on data tests of 50 test data were obtained the number of predictive positive sentiments of 31 data and negatives as much as 19 data, The results indicate that some features have already been met but still lack the maximum as often as an error message occurs when accessing the system and error in inserting the nim, and others.Based on the calculation matrix naive bayes on three attempts to split data with accuracy rate of 82.00 %, Precision of 72.73 %, and recall