Volume 14, Number 2
Making Medical Experts Fit4ner: Transforming Domain Knowledge through Machine Learning-Based Named Entity Recognition
Authors
Florian Freund, Philippe Tamla, Frederik Wilde and Matthias Hemmje, University of Hagen, Germany
Abstract
This study presents a comprehensive survey examining the criteria used by Machine Learning (ML) experts in selecting and comparing Names Entity Recognition (NER) frameworks. The survey revealed that while performance is a key criterion, expert opinions vary significantly, highlighting the need for a flexible system that considers various criteria alongside performance. Based on the survey results, a system was developed using the structured Nunamaker methodology to assist medical experts in both comparing NER frameworks and training ML-based NER models. The prototype, including its user interfaces, was qualitatively evaluated using the Cognitive Walkthrough method. The paper concludes with a summary and an outlook on future research.
Keywords
Natural Language Processing, Named Entity Recognition, Machine Learning, Cloud Computing, Medical Expert Systems, Clinical Decision Support