Analisis Komparasi Algoritma Machine Learning untuk Sentiment Analysis (Studi Kasus: Komentar YouTube “Kekerasan Seksual”)

Abstrak

Cases of sexual violence in the last decade have been rampant in Indonesia. Cases of sexual violence are increasingly exposed, along with the increasing use of social media. One of them is violence against women. Cases of sexual violence often cause various kinds of stigma in the community, so this study aims to determine the public's response to cases of sexual harassment using sentiment analysis. The data used is sourced from YouTube comments with the title "Kasus Bunuh Diri NW: Bripda Randy Tersangka, Penanganan Polisi Dikritik | Narasi Newsroom." The method used is Machine Learning algorithms such as the SVM algorithm, Naive Bayes, and Random Forest. The results of comparing the three Machine Learning algorithms, Random Forest, obtained the best accuracy rate of 78% compared to the other two algorithms in conducting sentiment analysis on YouTube comments about sexual harassment discussions.