Medical-Information-Based Data Mining Research

As Hospital Information Systems' development, hospitals have accumulated amountain of data. And utility, analysis, mining of information and patterns that hide in thosedata, can give doctors and hospital administrators some insights in diagnosis and hospitals'administration. The study of data mining in our country is just at its' beginning, and it is mainlyused telecom, finance and insurance fields. Moreover, most of the products are from abroad.However, Data mining technologies has been well studied and wildly used among manyindustries. And Healthcare industry is included. The Data mining technologies are not onlyused in Hospital management, but also used in assisting doctors' work. So, the study of datamining applications in healthcare industry is very valuable in our country.In this article, we focus on Hospital information data warehouse which is based on existedHospital information data, and explore valuable patterns which hide in the data collected in datawarehouse. So that doctors can use those patterns to diagnose patients or establish templates.Specifically speaking, this article can be divided into following four parts. First, organizemedical data by subjects according to data warehouse principle. Experts use those dimensionsto analyze various indicators, moreover, analyze, abstract and apply medical information fromdifferent angles. Second, construct medical data warehouse based subjects; then extract,transform, clear and load data from different Systems. Third, analyze and choose mining themesbased on medical information data warehouse, and describe problems which exist in medicalinformation mining. And construct data mining platform, analyze and evaluate the results. Inthe end, modify promote a new parameter to apriori algorithm. And use this algorithm to minemedical order data, and then find association rules which exist in medical order items.