Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Jurnal FASILKOM (teknologi inFormASi dan ILmu KOMputer)
Setiadi, Erik
Unknown Affiliation
Computer Science & IT Decision Sciences, Operations Research & Management

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Analisis Sentimen Kebijakan Makan Bergizi Gratis Menggunakan IndoBERT dan Machine Learning Sulistyo, Danang Arbian; Setiadi, Erik
JURNAL FASILKOM Vol. 15 No. 3 (2025): Jurnal FASILKOM (teknologi inFormASi dan ILmu KOMputer)
Publisher : Unversitas Muhammadiyah Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37859/jf.v15i3.10546

Abstract

Social media has increasingly become a primary channel for the public to express opinions and reactions toward government policies, including Indonesia’s Makan Bergizi Gratis (MBG) program. The large volume of online discussions and the diversity of publicperspectives highlight the importance of sentiment analysis to assess public acceptance and identify potential challenges in policy implementation. This study aims to analyze the distribution of public sentiment toward the MBG policy and to evaluate the performance of machine learning models for sentiment classification. The dataset consists of 12,389 Indonesian-language tweets collected from platform X. Sentiment labeling was performed automatically using a hybrid labeling approach that combines a domain-specific lexicon-based method with the IndoBERT deep learning model to improve contextual understanding. Sentiment classification was conducted using hybrid features, including TF-IDF trigrams, IndoBERT embeddings, and lexicon-based features. Three classification modelsRandom Forest, XGBoost, and an Ensemble modelwere trained and evaluated on a SMOTE-balanced dataset to address class imbalance. The results reveal that public sentiment toward the MBG policy is predominantly negative (68.6%), followed by positive (19.5%) and neutral (11.9%) sentiments. Among the evaluated models, Random Forest achieved the best performance, obtaining an F1-score of 0.9383 using K-Fold cross-validation and 0.9363 on the final test set. This study concludes that the proposed hybrid approach is effective and reliable for classifying public sentiment toward government policies in the Indonesian language.
Co-Authors Sulistyo, Danang Arbian