Research On Distributed Incremental Mining Of Epidemic Data

With the continuous development of the information construction for hospital,data mining technology has been widely used in the medical field. Doctors can getbetter treatment plan and save medical costs in treatment process for epidemic withthe help of data mining technology. The existing conclusions need to be improvedwith the addition of data. In current distributed Medical collaborative model,epidemic data are widely distributed and limited. It is important to find the rightincremental mining methods.For association rules in clinical pathway and decision tree in aided diagnosissystem, two issues are existed in recent research as following.[i] On incremental mining for association rules, recent methods scan old datamany times. But the epidemic data is so much that the methods needs to consumemuch time in scanning the old data.[ii] On incremental mining for decision tree, redundant data has a great influenceon building decision tree and the old decision tree is not good used.This paper concluded and analyzed the related research recently, established aglobal site based on current Distributed Medical collaborative model for data mining.On incremental mining for association rules, this paper proposed a new mining policyto reduce the number of scanning old data and proposed new pruning strategy toreduce the number of candidate itemsets; on incremental mining for decision tree, thispaper used old decision tree to remove the redundant part of incremental data on localsites and adjusted the old decision tree to improve the efficiency of getting the newdecision tree. This paper proved the algorithm feasibility with the experiment,summarized the work and contributes of the work, and proposed to the future work.