Putra, Brahmayudha Erlangga (2026) Analisis Persepsi Publik terhadap Program Makan Bergizi Gratis pada Platform X melalui Pendekatan Sentiment-Aware Topic Modeling Menggunakan IndoBERTweet dan Embedding Multilingual-E5. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003221084-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003221084-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy |
Abstract
Program Makan Bergizi Gratis (MBG) merupakan kebijakan pemerintah untuk mengatasi permasalahan gizi di Indonesia, namun pelaksanaannya memicu perdebatan dan polarisasi opini di platform X sehingga diperlukan analisis yang mampu mengungkap kecenderungan sentimen sekaligus isu yang mendasarinya. Penelitian ini bertujuan menganalisis persepsi publik terhadap Program MBG melalui pemetaan dimensi afektif dan kognitif masyarakat. Data penelitian berupa unggahan pada platform X yang dikumpulkan melalui crawling selama periode 6 Januari 2025 hingga 30 April 2026. Pendekatan yang digunakan adalah sentiment-aware topic modeling yang mengintegrasikan analisis sentimen menggunakan IndoBERTweet dengan IndoBERT sebagai baseline serta topic modeling menggunakan BERTopic dengan embedding Multilingual-E5 dan dibandingkan dengan BERTopic menggunakan embedding SBERT sebagai baseline. Hasil penelitian menunjukkan IndoBERTweet menghasilkan kinerja terbaik dengan accuracy sebesar 94,20% dan F1-score sebesar 93,11%, serta mengidentifikasi dominasi sentimen negatif sebesar 70,30% dibandingkan sentimen positif sebesar 29,70%. BERTopic dengan embedding Multilingual-E5 menghasilkan kualitas topik terbaik pada sentimen negatif dengan C_v score sebesar 0,6300, sedangkan BERTopic dengan embedding SBERT menghasilkan kualitas topik terbaik pada sentimen positif dengan C_v score sebesar 0,7546. Integrasi analisis sentimen dan topic modeling mengungkap dua irisan diskursus utama, yaitu esensi program dan aspek makanan, serta menunjukkan divergensi opini antara kedua kelompok sentimen. Sentimen positif didominasi narasi mengenai manfaat dan keberlanjutan program, sedangkan sentimen negatif didominasi kritik terhadap implementasi, tata kelola anggaran, kualitas makanan, dan transparansi pemerintah. Hasil penelitian menunjukkan bahwa penerimaan publik terhadap Program MBG tidak hanya dipengaruhi oleh tujuan program, tetapi juga oleh kualitas implementasi, akuntabilitas, dan transparansi penyelenggaraannya.
=======================================================================================================================================
The Free Nutritious Meal (MBG) Program is a government policy introduced to address nutritional problems in Indonesia. However, its implementation has triggered public debate and opinion polarization on platform X, highlighting the need for an analytical approach capable of identifying both sentiment tendencies and the underlying issues. This study aims to analyze public perceptions of the MBG Program by examining the affective and cognitive dimensions of public opinion. The data consisted of posts collected from platform X through a crawling process during the period from January 6, 2025, to April 30, 2026. A sentiment-aware topic modeling approach was employed by integrating sentiment analysis using IndoBERTweet with IndoBERT as the baseline, and topic modeling using BERTopic with Multilingual-E5 embeddings, which was compared with BERTopic using SBERT embeddings as the baseline. The results show that IndoBERTweet achieved the best performance with an accuracy of 94.20% and an F1-score of 93.11%, identifying 70.30% negative sentiment and 29.70% positive sentiment. BERTopic with Multilingual-E5 embeddings produced the highest topic quality for negative sentiment with a C_v score of 0.6300, while BERTopic with SBERT embeddings achieved the highest topic quality for positive sentiment with a C_v score of 0.7546. The integration of sentiment analysis and topic modeling revealed two overlapping discourse themes, namely the essence of the program and food-related aspects, while also highlighting divergent opinions between the two sentiment groups. Positive sentiment was dominated by narratives emphasizing the program's benefits and sustainability, whereas negative sentiment mainly focused on criticism regarding implementation, budget governance, food quality, and government transparency. These findings indicate that public acceptance of the MBG Program is influenced not only by its intended objectives but also by the quality of its implementation, accountability, and transparency.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Analisis Sentimen, IndoBERTweet, Makan Bergizi Gratis, Multilingual-E5, Topic Modeling, Free Nutritious Meal, IndoBERTweet, Multilingual-E5, Sentiment Analysis, Topic Modeling |
| Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Brahmayudha Erlangga Putra |
| Date Deposited: | 30 Jul 2026 01:45 |
| Last Modified: | 30 Jul 2026 01:45 |
| URI: | http://repository.its.ac.id/id/eprint/140179 |
Actions (login required)
 |
View Item |