The Application Of Data Mining In Data Analysis On Endocrine Hormone

With the progress of science and technology, more and more detecting methods are used for aided diagnosis. Meanwhile, a great deal of medical data is available. And how to find out the relation between factors and diseases from the data has become one of the research hotspots. In this paper, the endocrine hormone system and data mining are fully investigated, and statistical method is used to analyze the data first. And then, the character of medical data is analysed, based on the basic theories and realization methods of association rules, acquisition of valid association rules is studied.In this paper, firstly some basic knowledge of the endocrine hormone system and medical application of data mining are introduced. In the third chapter, through logistic regression analysis on data, some results are acquired first. And then in the forth chapter, at the basis of research on traditional mining algorithm for association rule, an improved algorithm for association rules with item constraints is proposed. The experiment result shows that the mining performance of the improved algorithm is better than the traditional Apriori algorithm. Meanwhile, for the disposal of incremental data, an improved incremental updating algorithm was used to avoid the re-mining of all data. In the fifth chapter, based on review of character of actual endocrine hormone data, some key components of association rules acquisition are studied, this part is one of the most important parts in this paper. Firstly, data discretization, neglect of vacant value and attribute transformation are used in the preprocessing of data to transform the original database into transaction database which is fit for mining. And then, the improved algorithm for association rules with item constraints proposed in the forth chapter was used to acquire the association rule from the transaction database. And then all rules were evaluated from the subjective and objective aspects to acquire the final valid rules, finally the results are compared with the results of logistic regression analysis in third chapter. At last we design and realize an information system for aided diagnosis.