Research On Electronic Medical Record Retrieval Technology Based On User’s Intentional Analysis

With the development of clinical medicine and the extensive application of medical information technology,it is more important to help to predict the condition in clinical medicine through the analysis of the correct diagnosis and treatment process.At the same time,largescale application and promotion of electronic medical records provided an important data base for subsequent analysis.As a key technology for large-scale information field,the application of information retrieval in the Internet has already been very mature.However,due to the particularity of the expression of medical information and the complexity of the relationship between the content,the information retrieval technology of medical is still in the initial stage.To cure the above problems,this paper achieved an understanding of the user’s query intent on the basis of dealing with the Chinese on electronic medical records,and retrieved the data by the means of graph retrieval and then diversify the results of the query.Finally,we achieved the objective of the auxiliary doctor analysis and diagnosed with the effective use of a large amount of data stored in electronic medical records,and improved the level and efficiency of medical services for doctors and medical staffs.This paper analyzed the structural characteristics of Chinese electronic medical records,and the semantics of electronic medical records,and extracted the medical entities and relationships of entities,laid the foundation for the follow-up study.In the aspect of analyzing the user’s query intent,this paper used the density-based clustering algorithm to analyze the sub-intent of the user’s query after clustering the historical data of user’s query to solve the breadth and ambiguity of the query.When consulting medical professional vocabulary,a method based on information entropy was proposed to calculate the similarity degree and concept relevance of information entropy.The calculation model of conceptual semantic similarity was obtained to identify the intention in medical concept.The medical concepts were classified by naive Bayesian classification primarily.The concept of information entropy calculated according to the concept of classification probability and the classification quality assessment function.Then the similarity degree of concept feature was obtained by entropy,and the conceptual relevance was obtained by analyzing the concept feature information.Finally,we used the weighted method to get the final semantic similarity of the first two values.Experiments showed that the algorithm is closer to the experience of expert evaluation in the medical field than the traditional algorithm,which can improve the accuracy of similarity calculation and can match the user’s intention more excellently or effectively.In order to obtain a better search result and search more effectively in electronic medical records,this paper proposed to express the structure of electronic data records with the graph structure.We put forward some improvements according to the similarity between the relationship between entities and attributes of the electronic medical record and topology of the graph.First,we graph structured the electronic medical record data and the retrieval results.Second,after combining the EMRSearch algorithm with the EMR-Tree index and introducing the new Upper bound model,we used the result to match the user’s intention.Finally,the Sort Diversity algorithm was used to reorder the matching result sets.Experimental results showed that this method can not only improve the efficiency of user’s retrieval,but also improve the user’s search satisfaction.