Analysis The Related Factors Of Diabetes Mellitus In Electronic Health Record With Data Mining

With the rapid development of database technology，it has infiltrated to a variety ofareas，including medicine．In order to solve the problem of sharing medical informationbetween the medical institutions, Many Medical information centers have set upRegional Health Information platform which is based on EHR. As those systemapplications and continuous development，the data quantity in the database rapidinflation, the database scale expands gradually and the complex degree increases day byday. However, despite the accumulation of a large number of EHR data, the real valueof these data has rarely been extracted and used in decision-making. So, it is meaningfulthat using effective method to extract useful information from EHR data. And diabetesmellitus is a disease that many people have been suffered. If we can extract some usefulinformation from the EHR system about the diabetes mellitus, we will put this usefulinformation into the health decision guide.So in this paper, data mining technology is used in analyzing the EHR data，andsome knowledge about diabetes mellitus is extracted from it. The main contributions areas follows：(1) The classification of mining technology is chose to create the Classificationmodel of diabetes mellitus patient based on EHR data. Tow data mining algorithms arestudied in this project. Finally, some classified rules are extracted from the model, andare used in predicting the high danger person who would be suffered with diabetes.(2)The personal’s EHR past medical history information is analyzed withassociation rule mining technology. And some association rules about the relationshipbetween diabetes and other diseases are found from the EHR data.(3)In order to create a system for diabetes decision, a data warehouse of EHR isdesigned based on the Regional Health Information platform database including theconceptual model, logic model and multi-dimensional model.