Analisis Sentimen Terhadap Permendikbud Ristek Nomor 30 Tahun 2021 pada Media Sosia Twitter Menggunakan Metode Lexicon-Based dan Multinomial Naïve Bayes

Abstract

Regulation of the Minister of Education, Culture, Research, and Technology (Permendikbud Ristek) Number 30 of 2021 was launched as a form of government efforts in the context of preventing and handling sexual violence in universities. However, it turns out that this regulation has generated various reactions from the community, most of them support it while others reject the ratification of this regulation. Technological developments that occur today encourage people to write their opinions on social media, one of which is Twitter. Tweets discussing this rule can be used to gauge public sentiment. However, considering the number of tweets, the classification process will be difficult to do manually, so it requires a computational system that can automatically classify the sentiments of the existing tweets. From these problems, a system is designed to perform sentiment analysis using the lexicon-based method and Multinomial Naïve Bayes. The results of this sentiment measurement can be useful as data analysis material for the Ministry of Education and Culture, Research and Technology in making decisions regarding this rule. The purpose of this research is to measure the value of accuracy, precision, recall, and f-measure in sentiment analysis using lexicon-based and Multinomial Naïve Bayes methods. The measurement results obtained using a dataset of 470 data are the accuracy value of 71.28%, precision of 70.10%, recall of 78%, and f-measure value of 74.29%.