Research On Semantic Similarity Algorithm And Realization Of Medical Term Service Platform

The rapid development of information technology to promote the hospital information construction,And medical information sharing is the focus of the current construction.As we all know,the standardization of terminology is the basis of information sharing,only make the terminology achieve unity from the bottom of the terminology can truly achieve the establishment of practical sharing of regional health information system.In recent years,the government has developed and published a series of standard specifications.But the lack of relevance between the terms of the standard make it not fully integrated and unified.In addition,terminology standardization is not just to develop standards,but also need to establish a platform.The establishment of medical terminology platform not only play the role of standard development and maintenance,but also play a role in promoting the application of standards,It is the underlying basis of regional health information platform.In this paper,through the study of foreign terminology service system,learn from foreign mature clinical medicine system SNOMED CT structure,improve its constituent components to build a online terminology service system which conform chinese medicine clinical term rules.The main work of this paper is as follows:(1)A computational method of semantic similarity of ontology in biomedical field based on Web search and SVM classifier is designed.This paper presents a method for calculating semantic similarity by combining the number of pages returned by two medical terms in the Baidu search platform and SVM classifier.By searching the terms P,Q and P AND Q respectively,the corresponding number of web pages are obtained.Five kinds of calculation methods are used to search for the features of the two terms P and Q,and the SVM classifier is used to obtain the corresponding F-score to evaluate the robustness of these methods,and finally a best classification model is obtained.Based on the vocabulary in SNOMED CT,the optimal correlation coefficient of the method is 0.802 with the medical expert score.The model can well calculate the semantic similarity of different ontology network vocabulary,and provide the basis for the new term classification of terminology service platform construction.(2)Establish a standardized medical terminology sharing service platform.This paper constructs a standardized medical terminology sharing service platform by using the method of SNOMED CT,which is a popular clinical term in clinical practice,and provides various functions related to the term.In order to strengthen its applicability,the platform also increased the medical literature subscriptions and annotations,personal terms space and other personalized functions.The platform includes the PC version and the mobile client version,which greatly covers the user population.In this paper,through the construction of the term service platform,we have created a comprehensive hardware and software environment to provide terminology services,around the terminology of the development,modification,coding,addition relationship,query and retrieval of the terms,the use of terms,the terms of the entire life cycle Management,and ultimately obtain the formation of a set of standardized clinical terminology system.And provide basic services for the follow-up of domestic medical information,medical data mining,intelligent medical,decision-making medical services.