Design And Implementation Of Event Evolution Graph Constructing System Based On Electronic Medical Records

With the improvement of the medical information construction level,medical institutions can record and store medical electronic medical records efficiently and conveniently.Electronic medical records record the diagnosis and treatment process of patients.It is the result of accumulated experience in the diagnosis and treatment of medical institutions,which contains a lot of rich and valuable medical knowledge.With the development of medical and health big data,the intelligent construction of electronic medical records has attracted attention.In recent years,the appearance of knowledge graph has greatly promoted the development of cognitive intelligence in the field of artificial intelligence.The emergence of the event evolution graph,which focuses on the events and the relation between events,helps the machine to learn the logic between events and master the law of development and evolution of events.The process of diagnosis and treatment recorded in electronic medical records can be abstractly summarized as medical events.Constructing event evolution graph in medical field based on the electronic medical records data can help to understand the law of development and evolution of disease,and then enables the medical knowledge precipitated in electronic medical records to better serve the diagnosis and treatment of diseases.This paper studies the problem of constructing event evolution graph based on Chinese electronic medical records.Most of the contents in Chinese electronic medical records are unstructured natural language texts.This paper aims to extract the medical events in Chinese electronic medical records and mine the relationship between medical events through natural language processing technology,and then constructing event evolution graph in medical field.On the basis of the above algorithm on event evolution graph constructing,this paper designs and implements an event evolution graph constructing system based on electronic medical records.The work of this paper includes the following aspects:1)This paper defines the event representation for Chinese electronic medical records,and uses a two-stage event extraction method to extract medical events from Chinese electronic medical records.In the first stage,the triggers of medical events are extracted,and in the second stage,the event arguments of medical events are extracted.2)Summarize the explicit expression templates of causality in Chinese electronic medical records,extract text fragments with causality in Chinese electronic medical records based on the pattern rule matching,and then extract the causal event pairs.3)Conduct demand analysis of the medical event evolution graph constructing system,design the overall structure and functional modules of the system in detail on the basis of demand analysis,and finally realize an event evolution graph constructing system based on electronic medical record.