Medizinische Universität Graz - Research portal
Selected Publication:
SHR Neuro Cancer Cardio Lipid Metab Microb
Kugic, A; Schulz, S; Kreuzthaler, M.
Domain Shift in Part-of-Speech Tagging.
Stud Health Technol Inform. 2025; 327: 896-897.
Doi: 10.3233/SHTI250497
PubMed
FullText
FullText_MUG
- Leading authors Med Uni Graz
- Kugic Amila
- Co-authors Med Uni Graz
- Kreuzthaler Markus Eduard
- Schulz Stefan
- Altmetrics:
- Dimensions Citations:
- Plum Analytics:
- Scite (citation analytics):
- Abstract:
- This study highlights domain shift in dataset distributions that impact machine learning performance in clinical natural language processing, analyzing linguistic differences across clinical narratives, biomedical abstracts, and news articles in English using part-of-speech (POS) tag distributions. Results indicate significant variations in POS tag occurrences, with undefined tags more frequent in clinical datasets, emphasizing the need for specialized tools and improved domain adaptation techniques to address these challenges.
- Find related publications in this database (using NLM MeSH Indexing)
- Natural Language Processing - administration & dosage
- Machine Learning - administration & dosage
- Humans - administration & dosage