Medical Adverse Events Classification for Domain Knowledge Extraction

摘要

In the viewpoint of sharing knowledge, users need to input keywords before the agent retrievals the related information from the Internet. At the same time, the traditional method ignored the true meaning of the terms. However, the semantic web has been created to improve the disadvantage. Ontology is the fundamental element of the semantic web that is a kind of knowledge presentation and can present the variety of property for concept.According to the regulation of Department of Health, Executive Yuan, R.O.C, the hospital has implemented the new Hospital Accreditation from 2005. Furthermore, the Hospital Accreditation evolves the variety of items of accreditation. Medical adverse event reporting is a kind of the Hospital Accreditation. The system which can deal with the medical adverse event reporting is operated by human at present. None the less, the system caused some questions, such as inconsistent knowledge and human negligence. The research proposed a new framework which can replace the human automatically. We get concepts through extraction of information and analyze sentences through classification of events. After that we analyze those classes which can construct variable and dynamic ontology in medical adverse event reporting.