Application Of Association Rules Mining In Disease Data Processing

With the computer science and technology rapid development and wide application, the amount of data generated in all medical disciplines were rapidly increasing. In order to find valuable knowledge and rules from the huge amounts of data, statistical, database, artificial intelligence and neural network technology were combined to data mining methods to solve this problem. The application for association rule mining was an important research topic in the field of data mining. At present, the use of association rules mining to discovery knowledge had become the focus of attention of various disciplines. It was important to find out the intrinsic association rules from a large number medical data by use of the association rule mining technology, which could not only provide an effective basis for disease monitoring, evaluation of drug treatments and disease prevention in clinical practice, but also was a new research content.The data mining of medical data, which was privacy, polymorphism, no integrity, timeliness and redundancy features, was analysed by using of classic Apriori algorithm of association rule mining. In order to find an useful correlation relationship or pattern among the itemsets in huge amounts of data, Apriori algorithm was used to find frequent itemsets of data from the database, and then frequent itemsets generated an strong association rules. The ultimate objective is to research and analyse the application of association rule mining in clinical disease surveillance, evaluation of drug treatments and disease prevention aspects. For example, through association rule mining analysed the data on complications of type 2 diabetes and drug treatment of liver disease patients, the reference of early prevention for the complications of type 2 diabetes and the decision making for the treatment of liver diseases could be obtained.Detailed analysis of this thesis in data mining and association rule mining baseing on the characteristics, the classic algorithm, including Apriori association rule mining algorithm, based on division algorithm and FP-tree algorithm frequency, were described and analyzed; the characteristics of medical data was analysed, association rules in the application of data related diseases was researched, and the data, which was obtained from three major complications (hyperlipidemia, hypertension, coronary heart disease) in the follow-up cases of type 2 diabetes and liver disease patients by using of the Apriori algorithm, was collected for data mining and data analysis. The reference was provided by the association rule mining results from the complications of type 2 diabetes for the early prediction of three main complications of type 2 diabetes. In clinical practice, patients with type 2 diabetes could be targeted to take on the three measures for examination and treatment of complications, in order to find the other symptoms induced by type 2 diabetes as early as possible and reduce the treatment costs and disease progression. Meanwhile, the decision-making basis of medication, which was offered by the results of liver disease association rule mining, could not only provide a rational choice of drugs , a best treatment of liver diseases drug programs and an improvement treatment effect and efficiency in clinical practice, but also could be used as drugs audit basis. It indicated that, the analysis of data mining on preventive health care and drug treatment and other aspects in the health field could be a worthy research direction.