EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Sentiment Analysis Model for the Free Lunch Program in Indonesia on Twitter (X) Based on Machine Learning

Amelia Tifany Dewi, Nur Alamsyah, Arnold Ropen Sinaga

Bulletin of Intelligent Machines and Algorithms · 31.03.2026 · 1 (3) · 88-93

DOI: 10.65780/bima.v1i3.18 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bulletin of Intelligent Machines and Algorithms, online erschienen am 31.03.2026. Verfasst von Amelia Tifany Dewi und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Social media has become a primary platform for the public to voice their opinions on various public policies, including the free lunch program initiated by the Indonesian government. This study aims to analyze public sentiment toward this program through the Twitter (X) platform by utilizing machine learning algorithms. Data collection was conducted from January 2025 to June 2025, with a total of 2,045 comments successfully gathered. Sentiment labeling was performed manually, and only positive and negative sentiments were considered. The data, in the form of relevant comments, were pre-processed and classified into positive and negative sentiments. Three algorithms used in this study are Support Vector Machine (SVM), Naïve Bayes, and Random Forest. Evaluation was performed using data splitting schemes of 70:30 and 80:20, along with 5-fold cross-validation. Unlike previous studies, which primarily focused on sentiment analysis of general social issues or specific topics without emphasizing public policy, this study specifically investigates the public's sentiment regarding a government policy (the free lunch program) and compares the performance of different machine learning models. The results of the study show that the Random Forest model outperformed SVM and Naïve Bayes, achieving an accuracy of 89.41% with a standard deviation of 0.0138. Meanwhile, SVM achieved an accuracy of 88.96% and Naïve Bayes 88.72%. These findings suggest that Random Forest is the most optimal and consistent model for sentiment analysis of public policies on social media.

Zitieren

APA

Dewi, A. T., Alamsyah, N., & Sinaga, A. R. (2026). Sentiment Analysis Model for the Free Lunch Program in Indonesia on Twitter (X) Based on Machine Learning. Bulletin of Intelligent Machines and Algorithms, 1(3), 88-93. https://doi.org/10.65780/bima.v1i3.18

Vancouver

Dewi AT, Alamsyah N, Sinaga AR. Sentiment Analysis Model for the Free Lunch Program in Indonesia on Twitter (X) Based on Machine Learning. Bulletin of Intelligent Machines and Algorithms. 2026;1(3):88-93. doi:10.65780/bima.v1i3.18

BibTeX
@article{Dewi2026,
  author = {Dewi, Amelia Tifany and Alamsyah, Nur and Sinaga, Arnold Ropen},
  title = {Sentiment Analysis Model for the Free Lunch Program in Indonesia on Twitter (X) Based on Machine Learning},
  journal = {Bulletin of Intelligent Machines and Algorithms},
  year = {2026},
  volume = {1},
  number = {3},
  pages = {88--93},
  doi = {10.65780/bima.v1i3.18}
}
Kostenlos starten →