The Application Of Data Mining In Medical Data Analysis

With the wild use of standard formatted digital case history and medical information system, the medical industry accumulated large amount of clinical data and management information. It is hard to find the inner relationship, or discover the pattern and the discipline by using the traditional analyses method. Data mining as a new data analyses method can be used in medical data analysis to extract the discipline, forecast trend, and discover the amazing pattern. It is very important for improving clinical diagnoses and hospital decision-making ability. The analyses method based on data mining is a new approach to boost the medical services quality and hospital's competitive advantages.Since the reformation of the medical system, the competition in medical market is more and more severe. The establishment of hospital total development strategy and the adjustment of market policy can directly affect a hospital's position and competency on medical market. Thus, the decision-making process is becoming more and more important. In addition, after several years of the first use of MHIS (Military hospital information system) in CAPF (China Army Police Force) hospitals, the large amount of cumulated clinical data and management information is a great foundation of scientific decision-making method.This article uses the original data collected from a CAPF hospital. According to surveys, we defined the analysis task and objective as following: identify the high profitable disease type to support the decision-making of a hospital's selection on characteristic disease type; identify the high occurrence area to support the decision-making of the marketing strategy; study the classic disease type to support the decision-making of the standard treatment. In addition, we use the classification method to analyze the follow-up survey data of diabetes mellitus.The clustering method is used in the research to identify the high profitable disease type which provides support for configuring the medical resources, selecting the characteristic disease type, adjusting the development strategy. Identify the main patient source area to support the formulation on medical marketing strategy. We use the association analysis method to detect the syndrome of diabetes mellitus, which provides reference for early prevention and cure of type 2 diabetes mellitus. We use the association analysis method to discover the relativity of uremia medicine, finds out the rule and warps of the therapy, provides the standard of uremia treatment and medicine audition. We use the C5.0 algorithm and BP neural network algorithm to build the type 2 diabetes mellitus forecast model which provides the reference for the clinical diagnoses.A part of this article's analysis result is the import source of the establishment of development strategy and adjustment of marketing tactical of the mentioned hospital.This article's analysis is based on the original data. It discusses the overall process from the data mining, data clearance, data pretreatment, exploring analysis, algorithm selection, model establishment to the evaluation of the MHIS. It is a good reference for the engineering implementation of medical data analysis. This article provides a better solution for the complex medical information analysis.This article is focused on the application. It combines the requirement to set the analysis task and objective, stands out the practical value of analysis results.