Research On Improved Association Rule Algorithm In Data Mining Of Chronic Diseases

Association rule mining is an important branch of data mining technology;its purpose is to discover the association between data items from a large amount of data.Since the form of the association rules is simple and easy to understand,the research and application of association rule technology has been flourishing.The number of patients with chronic disease in China is very large.To effectively use medical data of these patients and provide scientific basis for the prevention and control of chronic disease,medical data of hypertensive patients was selected to do data mining research in this paper.The association between signs of hypertension and cardiovascular risk level,and the association between hypertension and other chronic diseases were studied.Main contents of this paper were as follows:(1)Based on relevant domestic and foreign literature,data mining technology and its research status in chronic disease and other medical fields were analyzed.The problem of medical data analysis at the present stage was summarized.Main contents and route of the study were established.(2)Theories of data mining and association rules were expounded.The Apriori algorithm of association rule mining was studied emphatically.The bottleneck of its performance was analyzed,and the existing optimization methods were introduced,which broadened the train of thought for the improvement of this algorithm.(3)To improve the efficiency of Apriori algorithm,the following improvements were made: Clustering matrix strategy was introduced to compressively store transaction records,which can avoid scanning original transaction database;Pre-pruning strategy was introduced to generate fewer candidate item sets,which can avoid the large number of connections of frequent item sets;Constraints of chronic disease types were added,which can reduce frequent item sets and irrelevant rules.Finally,Matlab simulation experiments were conducted to verify the performance of this algorithm.It proved that the improved algorithm can effectively reduce the number of candidate item sets and improve operating efficiency.(4)A data mining scheme for medical data of the chronic diseases was designed.The improved Apriori algorithm was applied to the processing of the medical data of hypertensive patients.After preprocessing the data,setting the minimum support and confidence thresholds,limiting constraints and relativity,the association rule mining was conducted.Logistic regression analysis was used to explore the association between chronic diseases,and the results were compared with those generated by data mining.The results of two methods were consistent with each other and the validity of the experiment was proved.The experiment finally found out association rules which accorded with medical knowledge,which can accurately judge the cardiovascular risk level of the hypertensive patients and predict the complications of the chronic diseases.It provides a valuable reference for doctors to diagnose and a theoretical basis for the realization of automated diagnosis.(5)The data mining system for chronic diseases was developed and the improved Apriori algorithm was applied.Hidden knowledge behind the medical data of chronic disease can be discovered in this system,which can assist doctors in decision-making.This system is of great practical value.