Linguistic Rule-Based Methods for the Extraction of Medical Summaries to Benefit Patients Progression Tracking

摘要

Clinical narratives contain useful information that can complement the patient progress records which are obtained throughout the patient's medical and treatment duration. In order to understand the clinical narratives content, medical concepts that include events and temporal information should be performed. This study addresses this issue based on a linguistic rule-based approach which combines domain knowledge, extraction modules and temporal linker component. This is in contrast to the fundamentals adopted by the major works based on machine learning. The proposed work's performance is therefore evaluated against a machine learning based approach and a knowledge intensive approach. Results have shown its strength regardless of its different nature.