Research On Semantic Similarity And Reasoning Based On Unified Medical Language System

With the development of the information network, the Internet has penetratedinto various fields of social and economic life, which also includes the developmentof electronic medical records. Based on the results of the extracted electronicmedical record information, the general medical ontology or the Unified MedicalLanguage System can obtain users’ health status which is closely related to users’health knowledge. It is a shortcut to get knowledge representation of users’ health,which not only avoids the complicated hand-built knowledge base, but also inheritsseveral decades of accumulated medical knowledge that is available for thecomputer.The purpose of creating and distributing the Unified Medical Language Systemby U.S. National Library of Medicine is to improve the ability of computer programto understand the language of biomedical and health. Developers can use the UnifiedMedical Language System Knowledge Sources and associated software tools tocreate, retrieve and integrate biomedical and health data and information. TheUMLS Knowledge Source are multi-purpose and are used in systems that performdiverse functions involving information types such as patient records, informationretrieval systems and natural language processing.This paper makes a study on semantic similarity calculation of the UMLSMetathesaurus’s concepts and ontology reasoning of the UMLS semantic networkbased on the overall organizational structure of the Unified Medical LanguageSystem. The semantic similarity research establishes the English-Chinese medicalcontrolled vocabularies organized by using the unique identifier for concept in theMetathesaurus. This research maps the Chinese medical concepts into theMetathesaurus concepts in the UMLS, and then calculates the Chinese medicalconcepts using two different semantic similarity methods such as path-based methodand information-content-based method. The experimental result shows that using theMetathesaurus’s concept in the UMLS to calculate semantic similarity of theChinese medical field is feasible.By using OWL to represent the user’s personal health information in themedical field, people can make the knowledge with a standardized, easy sharingfeatures and the ability of logical reasoning. The ontology reasoning research mainlyfocuses on the reasoning that comes with all the semantic types and semanticrelations in the UMLS semantic network with the help of the framework of theSemantic Web. This paper uses OWL to describe the medical domain ontology, anduses Protégéwhich is an ontology building tool to achieve a consistency check and instance classification of the medical domain ontology.