Evaluasi IndoBERTweet untuk Klasifikasi Sentimen Ulasan Wisata melalui Inferensi Langsung Lintas Domain: Studi Kasus Kota Baubau
JURNAL INFORMATIKA · 05.06.2026 · 15 (1) · 60-72
DOI: 10.55340/jiu.v15i1.2683 ↗Kurz gefasst
Fachartikel in der Zeitschrift JURNAL INFORMATIKA, online erschienen am 05.06.2026. Verfasst von Nurul Hidayah und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Tourist reviews on digital platforms represent a valuable source of public perception data to support evidence-based tourism management. This study analyzes the capability of IndoBERTweet in classifying tourist review sentiment through a cross-domain direct inference approach without additional fine-tuning, with a case study on Benteng Keraton Buton and Pantai Nirwana in Baubau City. A total of 896 reviews were collected from Google Maps via web scraping and processed through case folding, cleaning, text normalization, and negation word concatenation to classify sentiment into positive, negative, and neutral classes. Evaluation using a confusion matrix yielded an accuracy of 0.67 and a weighted F1-Score of 0.71, with the best performance on the positive class (F1: 0.80) and lower performance on the negative (F1: 0.41) and neutral (F1: 0.44) classes. These findings indicate that IndoBERTweet has potential as an initial instrument for local tourism sentiment analysis in the absence of domain-specific labeled datasets, and the results may serve as a preliminary consideration for tourism management policy priorities in Baubau City; however, further performance improvement requires fine-tuning with labeled data from a relevant tourism domain.
Zitieren
APA
Hidayah, N., Dodiman, D., Asniati, A., & Darmin, L. O. D. P. (2026). Evaluasi IndoBERTweet untuk Klasifikasi Sentimen Ulasan Wisata melalui Inferensi Langsung Lintas Domain: Studi Kasus Kota Baubau. JURNAL INFORMATIKA, 15(1), 60-72. https://doi.org/10.55340/jiu.v15i1.2683
Vancouver
Hidayah N, Dodiman D, Asniati A, Darmin LODP. Evaluasi IndoBERTweet untuk Klasifikasi Sentimen Ulasan Wisata melalui Inferensi Langsung Lintas Domain: Studi Kasus Kota Baubau. JURNAL INFORMATIKA. 2026;15(1):60-72. doi:10.55340/jiu.v15i1.2683
BibTeX
@article{Hidayah2026,
author = {Hidayah, Nurul and Dodiman, Dodiman and Asniati, Asniati and Darmin, La Ode Dwiyan Pramono},
title = {Evaluasi IndoBERTweet untuk Klasifikasi Sentimen Ulasan Wisata melalui Inferensi Langsung Lintas Domain: Studi Kasus Kota Baubau},
journal = {JURNAL INFORMATIKA},
year = {2026},
volume = {15},
number = {1},
pages = {60--72},
doi = {10.55340/jiu.v15i1.2683}
}
EUVIMED