IPB University Logo

SCIENTIFIC REPOSITORY

IPB University Scientific Repository collects, disseminates, and provides persistent and reliable access to the research and scholarship of faculty, staff, and students at IPB University

AI Repository
 
Building and Categories


      View Item 
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Statistics and Data Sciences
      • View Item
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Statistics and Data Sciences
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Perbandingan Performa Algoritma Clustering Berbasis Partisi dan Kepadatan pada BERTopic Isu Makan Bergizi Gratis di Media Sosial X

      Thumbnail
      View/Open
      Cover (478.9Kb)
      Fulltext (4.938Mb)
      Lampiran (378.8Kb)
      Date
      2026
      Jenis/Type
      Skripsi
      Subtype
      Undergraduate Theses
      Author
      ADISECHA, RAIHAN SABILLAH
      Fitrianto, Anwar
      Rizki, Akbar
      Metadata
      Show full item record
      Abstract
      Perkembangan statistika dan sains data mendorong pengembangan topic modelling untuk mengidentifikasi tema utama dari data teks tidak terstruktur berskala besar, salah satunya BERTopic yang memanfaatkan transformer embeddings untuk menghasilkan topik yang lebih koheren. Penelitian ini menerapkan BERTopic pada 13.142 data opini publik terkait Program Makan Bergizi Gratis (MBG) di media sosial X periode 6 Januari 2025–6 Januari 2026, sekaligus membandingkan performa algoritma K-Means, K-Medoids, DBSCAN, dan HDBSCAN. Data diproses menggunakan embedding IndoSBERT-large, direduksi dengan UMAP, dikelompokkan dengan keempat algoritma yang dioptimasi menggunakan grid search, dan dievaluasi menggunakan coherence score sebagai metrik utama serta diversity score dan persentase noise sebagai metrik tambahan. Hasil menunjukkan algoritma berbasis kepadatan lebih unggul dibandingkan algoritma berbasis partisi, dengan DBSCAN sebagai yang terbaik (coherence score 0,711, diversity score 0,933, noise 0,495%), menegaskan efektivitasnya untuk BERTopic pada teks berbahasa Indonesia dalam mengungkap isu utama diskursus publik mengenai MBG.
       
      The advancement of statistics and data science has driven the development of topic modeling, particularly BERTopic, which leverages transformer embeddings to produce more coherent topics than traditional methods. This study applies BERTopic to 13,142 public opinion posts on the Free Nutritious Meal (Makan Bergizi Gratis, MBG) program from platform X, collected between 6 January 2025 and 6 January 2026, and compares K-Means, K-Medoids, DBSCAN, and HDBSCAN clustering algorithms. Data were processed using IndoSBERTlarge embeddings, reduced with UMAP, clustered using the four grid-search-optimized algorithms, and evaluated using coherence score as the primary metric, with diversity and noise percentage as supporting metrics. Results show densitybased algorithms outperform partition based ones, with DBSCAN achieving the best performance (coherence = 0.711, diversity = 0.933, noise = 0.495%), confirming its effectiveness for BERTopic on Indonesian-language text in uncovering major issues in public discourse on MBG.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/177594
      Collections
      • UF - Statistics and Data Sciences [166]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository