Claim Missing Document
Check
Articles

Found 1 Documents
Search

Comparative Analysis of Naive Bayes Model Performance in Hate Speech Detection in Media Social Twitter Muhammad Hadyan Baqi; Yuliant Sibaroni; Sri Suryani Prasetiyowati
JURIKOM (Jurnal Riset Komputer) Vol 10, No 1 (2023): Februari 2023
Publisher : STMIK Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/jurikom.v10i1.5493

Abstract

Twitter is a popular social media in Indonesia, and for some people, it is a place to find and disseminate information. Hate speech is aggressive behavior against individuals or groups such on race, gender, religion, nationality, ethnicity, sexual orientation, gender identity, or disability. In this study, hate speech is modeled using Naive Bayesian models, which consist of Multinomial, Bernoulli, and Gaussian Naïve Bayes Models. These methods were chosen because Naïve Bayes is a simple method but has good performance in the case of sentiment analysis. This research aims to get the method with the highest accuracy value in analyzing hate speech. Thus, the Naïve Bayes model can provide the best solution for hate speech problems. The process carried out in this study is to process all data which obtained from Twitter social media and then classify it using the Multinomial Naïve Bayes, Gaussian Naïve Bayes, and Bernoulli Naive Bayes models based on the classification of HS and non-HS sentiment categories.  In this study, to get the best accuracy, two different scenarios were used. The result of the analysis of the accuracy is 82.13% of the Multinomial Naïve Bayes model which is the best accuracy rate value compared with other models.