Research Of Mining Association Rules Based On The Mutiple Minimum Supports

With the development and advance of the society,more and more data has accuated in the database. How to find what they need from the vast amounts of data rules has become a problem what people have to face.It is very tedious for us to mining rules from s large amount of data, on the one hand, as in the real society, the data we get always have missed part of the value, it is very adverse for us to mining the rules,on the other hand,some condition attributes for decision making in the database that is unnecessary,when we mining the rules,we need not to cosider this condition,the existence of the redundant attributes partly to make rules mining more trouble.At the same time, for most of association rule mining using the single minimum support, this is not conducive to the actual application. Aiming at the problems above, this article will completion incomplete information system in rough set theory, attribute reduction combined with association rules mining,do research of mining association rules based on the mutiple minimum supports.The main research content is as follows:1. Simply introduced the theory of rough sets and association rules, and association rule mining problem research status at home and abroad, and carries on the corresponding research.2. On the basis of the original Apriori algorithm and combining with the modified golden ratio, it is concluded that the association rule mining algorithm which is based on the golden ratio improved algorithm,it solved the problem of the minimum support through to a single, at the same time, the late simulation experiments show that this algorithm can effectively improve the efficiency of mining association rules.3. Aiming at mining association rules based on the mutiple minimum supports,we improved the traditional algorithm in the paper.Firstly,sampling from the complete data which is in the incomplete information systems,firstly,we need to set a higher support threshold,then we using association rule mining algorithm based on the golden ratio after sampling of data mining,it is concluded that the frequent itemsets.After that completion and attributes reduction of incomplete data based on the frequent itemsets and some rules,then we can get a new information system. At last,we using the association rule mining algorithm based on the golden ratio to conclud the association rules in incomplete information system.Practical feasible late simulation experiment proves that compared with oth er algorithms,the combination algorithm which is proposed in this paper combin ation algorithm significantly improves the mining association rules in incomplet e information system efficiency.Though mining association rules in the complete information system has many research results,mining association rules in incomplete information system research is less.At the same time,this paper completion of the incomplete information system and reduct the attribute based on the mining association rules,this provides a possible research direction.