The Application Of Association Rules In Analysis Of Medical Insurance Data

.Modern medical insurance data management system has gradually become an important part of the integrated medical system which has been widely used.A huge amount of health insurance data which contains ample information is produced every day.These health insurance data contains a wealth of information on these data Effective analysis which can dig out the valuable decision-making information plays an important role in promoting health insurance reform.But it's hard to be coped with and analyzed because there exists noises,incompleteness and redundancy in the data.At the same time,there are some shortcomings among the methods that are currently analyzed by positive association rules,and cannot provide more comprehensive information for decision makers.Therefore,this paper first prepares the medical insurance data from a hospital,and then proposes an improved association rule mining algorithm MMS_FP based on multi-support and an improved positive and negative association rule based on two-level multi-support Mining algorithm 2LFP_inFS_FS.Finally,these improved algorithms were used to analyze the data of cardiovascular and cerebrovascular diseases,diabetes mellitus data and rheumatoid arthritis.Some positive and negative association rules of medication behavior and some diseases were analyzed between drugs and diseases.We analyzed them and the results we have got have important reference value for medical insurance reform,like the following:1.Propose an improved association rule mining algorithm based on multi-support——MMS_FPIn this paper,we propose a new association rule mining algorithm based on multi-support,MMS_FP,which takes the practical problems of different probability and frequency of each transaction into account,sets different degrees of support for each item in the data set,MMS model solves the problem of redundancy of frequent itemsets and uses the FP-Growth algorithm to achieve,so the improved algorithm runs several times faster than MSapriori,and can find more valuable item sets,to provide better support for the analysis of association rules.2.Propose an improved positive and negative association rule mining algorithm based on two levels of multi-support ——2LFP_inFS_FS2LS model is proposed to mine the non-frequent itemsets,but the model still sets single-level support for the entire itemsets with two levels support,neglecting the problem of different probability and frequency of each transaction in the itemsets.Therefore,we integrate the 2LS model with the XMMS model,that is,we set two levels of multi-support for each item in the itemset and propose a new positive-negative association rule mining based on two levels of multi-support The algorithm 2LFP_inFS_FS is implemented by using the FP-Growth algorithm while mining frequent and infrequent itemsets at the same time.The experimental results show that this algorithm is more efficient and then obtains the positive and negative association rules simultaneously through the PNARC model.3.Application of association rules in three disease data analysisFor medical insurance data is noisy,incomplete and redundant features,this paper does a lot of data pre-processing work: some of the data in the vacancy value to ignore elements,fill in and delete the way,some of the information in the data semantics Consistent operation of a number of discrete data in the data were processed by the rules,and finally selected the more attention to today's three kinds of diseases in society: cardiovascular and cerebrovascular disease,diabetes and rheumatoid arthritis data were positive and negative Rule mining and analysis,excavated a few meaningful association rules,such as cardiovascular and cerebrovascular diseases: West and with atropine sulfate match,access to information that the West dissolved in atropine sulfate precipitation phenomenon,so This combined medication is unreasonable,does not meet the medical requirements;and captopril and digoxin,captopril will increase the incidence of digitalis poisoning can not be combined with medication,in line with medical regulations and common sense.Similarly,we have found many rules of value in the treatment of diabetes mellifluous and rheumatoid arthritis.The results have important reference value for the diagnosis and treatment of common diseases and the revision of rational medical prescriptions.