Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal INOVTEK Polbeng - Seri Informatika
Rofi Tulus
YARSI University
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Comparative Evaluation of Hand-Engineered Features for CRF-Based Named Entity Recognition in Indonesian Clinical Research Articles Nurmaya Maya; Rofi Tulus; Nova Eka Diana; Herika Hayurani; Sri Puji Utami Atmoko; Yurika Sandra
Journal of Innovation and Technology Polbeng Series on Informatics (INOVTEK Polbeng - Seri Informatika) Vol. 11 No. 3 (2026): August (Inpress)
Publisher : P3M Politeknik Negeri Bengkalis

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35314/y1v81j78

Abstract

Extracting clinical entities such as signs, symptoms, diagnoses, and treatments from Indonesian clinical research makes clinical information more accessible to readers with varying levels of medical knowledge. Conditional Random Fields (CRFs) combined with hand-engineered features have proved effective for health domain named entity recognition (NER). However, studies comparing hand-engineered feature combinations for clinical NER in Indonesian clinical research articles remain limited. This study compares five hand-engineered feature combinations using a CRF model to identify the most effective combination for recognizing seven clinical entity types, using a newly annotated corpus of 29 Indonesian internal medicine research articles. The dataset was split into 80% training and 20% testing data. Model performance was evaluated using weighted-average and macro-average precision, recall, and F1-score, reported as mean ± SD across five random states (5, 10, 15, 20, 25) and training epochs (50, 100, 150). The combination of contextual, prefix and suffix, and orthographic features achieved the best weighted F1-score of 78.90% at Epoch 50, though macro-F1 was considerably lower at 64.32%, reflecting weaker performance on minority entity types. This gap indicates a need for further corpus expansion and data balancing.
Co-Authors Herika Hayurani Nova Eka Diana Nurmaya Maya Sri Puji Utami Atmoko Yurika Sandra