Hybrid Subword: Character Representation for Robust Sentiment Classification on Multilingual and Code-Mixed Indonesian Text
Journal of Electrical, Computer and Informatics · 31.03.2026 · 2 (1) · 11-24
DOI: 10.70062/jeci.v2i1.276 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Electrical, Computer and Informatics, online erschienen am 31.03.2026. Verfasst von Wendy Sarasjati und 2 weiteren Autor:innen.
Abstract (Originalsprache)
User-generated Indonesian text frequently exhibits code-mixing with English (“Indonglish”), informal spelling, elongation, and keyboard typos. These phenomena break subword tokeniza tion assumptions and may degrade multilingual Transformer performance in deployment. This paper studies a hybrid representation that fuses XLM-R sentence features with a character-level CharCNN branch designed to capture orthographic patterns and mitigate character noise. We evaluate (i) a standard XLM-R fine-tuning baseline, (ii) an ablation that removes the character branch (NusaX only), and (iii) the proposed hybrid model on two datasets: NusaX-Senti (12 regional languages) and Indonglish (Indonesian–English code-mixed sentiment). Beyond clean test performance, we introduce a controlled robustness protocol by injecting character-level perturbations with probability p=0.18 and measuring performance drop. Results show that the XLM-R baseline achieves the best clean Macro-F1 on both datasets, while the hybrid model substantially improves robustness on Indonglish by reducing Macro-F1 drop from 0.030 to 0.007 under noise. We analyze common error confusions and discuss when character-aware features help or harm across languages.
Zitieren
APA
Sarasjati, W., Ansor, B., & Windasari, A. (2026). Hybrid Subword: Character Representation for Robust Sentiment Classification on Multilingual and Code-Mixed Indonesian Text. Journal of Electrical, Computer and Informatics, 2(1), 11-24. https://doi.org/10.70062/jeci.v2i1.276
Vancouver
Sarasjati W, Ansor B, Windasari A. Hybrid Subword: Character Representation for Robust Sentiment Classification on Multilingual and Code-Mixed Indonesian Text. Journal of Electrical, Computer and Informatics. 2026;2(1):11-24. doi:10.70062/jeci.v2i1.276
BibTeX
@article{Sarasjati2026,
author = {Sarasjati, Wendy and Ansor, Basirudin and Windasari, Anggry},
title = {Hybrid Subword: Character Representation for Robust Sentiment Classification on Multilingual and Code-Mixed Indonesian Text},
journal = {Journal of Electrical, Computer and Informatics},
year = {2026},
volume = {2},
number = {1},
pages = {11--24},
doi = {10.70062/jeci.v2i1.276}
}
EUVIMED