ANALISIS SENTIMEN MASYARAKAT TERHADAP KEBIJAKAN PEMBERANTASAN JUDI ONLINE DI INDONESIA MENGGUNAKAN NATURAL LANGUAGE PROCESSING PADA MEDIA SOSIAL

Abstract

The rapid growth of online gambling in Indonesia has become a serious social problem, prompting the government to implement various eradication policies. This study aims to analyze public sentiment toward Indonesia's online gambling eradication policy using Natural Language Processing (NLP) techniques on social media data collected from YouTube. A total of 237 comments were gathered and processed through preprocessing stages including cleaning, normalization, stopword removal, and stemming using the Sastrawi library. Sentiment labeling was performed using a weighted lexicon-based approach with 600+ sentiment words. Classification was conducted using three models—Random Forest, Linear SVM, and Logistic Regression—with SMOTE applied for class balancing and 5-fold cross-validation for robustness evaluation. The best model, Linear SVM, achieved an accuracy of 97.92% and a CV score of 97.56%. Results showed that 57.4% of public sentiment was neutral, 28.3% positive, and 14.3% negative, indicating that the majority of the public responds to this issue informationally. This study demonstrates that NLP-based sentiment analysis is an effective tool for evaluating public perception of digital policy in Indonesia.