IndoVetBERT: A Domain-Adaptive Transformer for Indonesian Veterinary Clinical Text Classification
Year: 2026
Type: article
Abstract: Medical Health Records (MHRs) contain structured and unstructured data, such as textual information that supports the diagnostic process, informs clinical decision-making, and improves healthcare efficiency. This textual information is often overlooked compared to structured data due to its high dimensionality. Natural Language Processing (NLP) is a powerful tool for analyzing text data and classifying clinical text. Veterinary MHRs are more difficult to understand than human MHRs due to the limited number of datasets, the mixed use of English and Latin, and the lack of a robust ontology standard. To overcome language and domain gaps in prior models, this study proposes IndoVetBERT, a transformer-based model for analyzing veterinary MHRs in Indonesia. Compared to baseline models mBERT, XLM-R, and IndoBERT, IndoVetBERT achieved an accuracy of 87%. The proposed model effectively handles veterinary clinical narratives and captures veterinary-specific terminology, clinical reasoning patterns, and diagnostic clues in veterinary records in Indonesia.
DOI: 10.48084/etasr.19231
Institutions: Universitas Gadjah Mada, Universitas Amikom Yogyakarta
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $431
OpenAlex ID: W7202246776