Embedding model for the Malagasy informal language
SYSTEM THEORY, CONTROL AND COMPUTING JOURNAL · 15.12.2025 · 5 (1) · 1-13
DOI: 10.52846/stccj.2025.5.1.64 ↗Kurz gefasst
Fachartikel in der Zeitschrift SYSTEM THEORY, CONTROL AND COMPUTING JOURNAL, online erschienen am 15.12.2025. Verfasst von Francis Rakotomalala und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Processing informal Malagasy language presents major challenges due to linguistic variations, abbreviations, and frequent code-switching in digital communication. This study proposes a text embedding model based on DistilBERT and XML-RoBERTa, specifically adapted to informal Malagasy. Through fine-tuning on custom corpora, we observe a gradual improvement in performance, with a significant reduction in loss function and lower perplexity, indicating a better understanding of linguistic structures. The evaluation shows that the generated embeddings effectively capture semantic similarities, even across varied formulations. DistilBERT outperforms XML-RoBERTa, demonstrating better generalization. These results highlight the importance of adapting language processing models to low-resource languages and open up new perspectives for applications in the automatic understanding of informal language.
Zitieren
APA
Rakotomalala, F., Hajalalaina, A. R., & Ravonimanantsoa, N. M. V. (2025). Embedding model for the Malagasy informal language. SYSTEM THEORY, CONTROL AND COMPUTING JOURNAL, 5(1), 1-13. https://doi.org/10.52846/stccj.2025.5.1.64
Vancouver
Rakotomalala F, Hajalalaina AR, Ravonimanantsoa NMV. Embedding model for the Malagasy informal language. SYSTEM THEORY, CONTROL AND COMPUTING JOURNAL. 2025;5(1):1-13. doi:10.52846/stccj.2025.5.1.64
BibTeX
@article{Rakotomalala2025,
author = {Rakotomalala, Francis and Hajalalaina, Aimé Richard and Ravonimanantsoa, Ndaohialy Manda Vy},
title = {Embedding model for the Malagasy informal language},
journal = {SYSTEM THEORY, CONTROL AND COMPUTING JOURNAL},
year = {2025},
volume = {5},
number = {1},
pages = {1--13},
doi = {10.52846/stccj.2025.5.1.64}
}
EUVIMED