APPLYING TEXT MINING FOR EVIDENCE-BASED POLICY: SENTIMENT AND TOPIC ANALYSIS OF INDONESIA’S RESEARCH FUNDING REFORM
Jurnal Statistika dan Aplikasinya · 30.06.2026 · 10 (1) · 39-53
DOI: 10.21009/jsa.10104 ↗Kurz gefasst
Fachartikel in der Zeitschrift Jurnal Statistika dan Aplikasinya, online erschienen am 30.06.2026. Verfasst von Rolando Gultom und 11 weiteren Autor:innen.
Abstract (Originalsprache)
A substantial portion of governance-relevant information in Indonesia is contained in unstructured textual artefacts such as meeting minutes and focus group discussion (FGD) notes, yet these materials remain underutilized in policy evaluation. Existing studies have not systematically examined how computational text analysis can extract policy insights from such documents. The purpose of this research is to evaluate the implementation dynamics of the RIIM funding scheme during its transition to SBK, and to generate evidence-based recommendations that support the development of a more adaptive and context-sensitive research funding framework. This study employs a text-mining approach combining exploratory word cloud visualization, lexicon-based sentiment analysis using the Bing lexicon, emotion analysis using the NRC Emotion Lexicon, and Latent Dirichlet Allocation (LDA) topic modeling. The analysis is conducted on a corpus derived from Focus Group Discussion (FGD) verbatim transcripts involving RIIM stakeholders, consisting of 13 pages and 4,188 words. The findings reveal four major issue clusters: administrative and contractual inconsistencies, ambiguity in output classification across research fields, uncertainty in financial and temporal transitions, and challenges in aligning performance evaluation with diverse research outputs. These findings inform concrete recommendations for enhancing procedural clarity, establishing explicit transitional provisions, strengthening output classification, and institutionalizing stakeholder feedback mechanisms. This study acknowledges limitations related to the use of English-based sentiment lexicons for Indonesian data and the reliance on summarized minutes rather than verbatim transcripts. The study’s originality lies in demonstrating how computational text analysis can systematically extract policy insights from unstructured governance documents, offering a novel evidence-based approach for refining Indonesia’s research funding instruments.
Zitieren
APA
Gultom, R., Lelono, R. A. A., Juhartono, Kusumawati, L., Styawan, D., Rino, Purnomo, S., Prihartono, A., Maiforlion, Firmansyah, M., Faishal, A., & Handoko, D. R. (2026). APPLYING TEXT MINING FOR EVIDENCE-BASED POLICY: SENTIMENT AND TOPIC ANALYSIS OF INDONESIA’S RESEARCH FUNDING REFORM. Jurnal Statistika dan Aplikasinya, 10(1), 39-53. https://doi.org/10.21009/jsa.10104
Vancouver
Gultom R, Lelono RAA, Juhartono, Kusumawati L, Styawan D, Rino, et al. APPLYING TEXT MINING FOR EVIDENCE-BASED POLICY: SENTIMENT AND TOPIC ANALYSIS OF INDONESIA’S RESEARCH FUNDING REFORM. Jurnal Statistika dan Aplikasinya. 2026;10(1):39-53. doi:10.21009/jsa.10104
BibTeX
@article{Gultom2026,
author = {Gultom, Rolando and Lelono, Raden Arthur Ario and Juhartono and Kusumawati, Lianna and Styawan, Dody and Rino and Purnomo, Setyo and Prihartono, Agus and Maiforlion and Firmansyah, Maman and Faishal, Akhmad and Handoko, Danang Rita},
title = {APPLYING TEXT MINING FOR EVIDENCE-BASED POLICY: SENTIMENT AND TOPIC ANALYSIS OF INDONESIA’S RESEARCH FUNDING REFORM},
journal = {Jurnal Statistika dan Aplikasinya},
year = {2026},
volume = {10},
number = {1},
pages = {39--53},
doi = {10.21009/jsa.10104}
}
EUVIMED