Health Check-up Data Warehouse And Relevant Data Mining

With the continuous development of information technology and more attention paid to healthy, there is a great deal of valuable data accumulated in the healthy database which has not yet been fully utilized. Therefore, how to make sufficient use of the existing data to better support doctors and managers to provide better medical examination to customers is the focus of this article.Firstly, as the historical data is increasing and the CIMING company expands year by year, the original database system is not only unable to load the sharp increase amount of data. but also not suitable for data analysis and for the manager to make strategic decision from the perspective of the whole enterprise, building data warehouse is the only way. Therefore, this paper builds a model of distributed data warehouse and describes the architecture design. logic model design and the method of data analysis.Secondly, in order to solve the issue that how to make a reasonable examination of the items set, this paper focuses on the research of the level weighted association rules. The weight design learns from the TF*IDF method in the field of traditional IR. and gives the formula of weight definition for each department. Combining with the weight normalized method, so that the weighted support is no more than 1, which is easier to understand. Finally, by comparing the weighted association rules algorithm and the traditional non-weighted association rules algorithm in the application of this article, proves that the weighted association rules algorithm is more suitable and completes the mission of making items set.Finally, in order to assiste doctor to recommend appropriate items set for each user, this paper have a research on association rules mining based on rough set and especially pay more attention to the core calculating method. Due to the pre-decision table is serious incompatible, and the incompatible items are all fact, so the original method is no longer suit in this paper. For this reason, this paper proposes a vertical-weighted calculating core method: according to incompatibility degree of the data, weights for the record vertically, delete the records whose right are lower than the given value, and then calculates the right of core property according to the right of the records, finally, decides the core property according to each right value and the given value. The experiments show that, the method has good flexibility and with the help of doctors, it realizes the project very good.Based on the database of CIMING, this paper uses the above mehods to carry out result of the data mining and achieve the objective of decision-making.