Research On Resident Health Records Data Clustering Analysis

In the past few decades, China’s Hospital Information System along with the continuous development of electronic information technology has been more mature. Widely used in hospital information systems and digital medical equipment, which resulting in hospital data storage is growing rapidly, large and very large databases were applied to store the massive hospital data. In fact, for most hospitals, processing data in the database is only confined to the routine transactions, which involved in add, delete, modify and retrieval. Lack of data integration and analysis, face the dramatically increases the database, showed the phenomenon of "data rich, knowledge poor". Resident health record is a branch of the development of hospital information system, mainly recorded by community residents health information of periodic physical examination, which is reflected in the following three aspects: facilitate the health administration grasp the individual’s health status, provide medical staff in patients with corresponding detailed historical data and provide useful information supporting health administrative departments making disease prevention and control programs and the development of health policies and regulations. Because the characteristics of medical data determines the complexity of its data mining, it failed to understand the health records stored in the database data, to a certain extent, valuable information resources are wasted. How to maximize use these valuable resident health records for them provide timely personal health status, clinical research, provide scientific guidance for diagnosis and treatment of disease, provide scientific guidance basis for health administrative departments to decision-making, has become a current resident health record system construction problem.In view of this situation, this paper analyzed the status of residents’ health record system, put forward to establish health record system based on data warehouse, and apply data mining clustering method for the analysis of residents’health record, data mining theory and medical information organic together, find the data changes regulation, summarize valuable summary.The main topics of the study are as follows:1. On data mining theory and methods, focusing on analysis of the clustering algorithm K-means, lists its advantages and disadvantages and make a detailed comparative, put forward an improved K-means algorithm, simulation results show that the improved algorithm in time-complexity and space-complexity have significantly improved;2. Based on "national basic public health service", studied our country residents’health record status, combined with the subject of research, extract diabetes associated portion of the data, conduct a clustering algorithm of K-means application.3. Construct a prototype system of residents’ health record, which can specific and visualization the study content, more understanding the residents’ health record characteristics and the application of data mining technology.