Research And Application Of Weighted Association Rules Algorithm Based On Cluster And Compression Matrix

Association rules mining is an important technology of data mining,which discovers previously unknown and interesting relationships among attributes in the large databases. To sovle the different importance and unbalance of individual items in database,this paper proposes an improved algorithm for mining the weighted association rules algorithm which based on cluster and compression matrix——CCMW algorithm.The works of this dissertation are as follows:(1)This paper discusses data mining techniques,and then emphasizes the basic concepts of cluster analysis and association rules,their ideas,and the representative algorithms——K-Means algorithms and Apriori algorithm.(2)The famous algorithm——MINWAL(O) alogrithm is thoroughly studied. Some problems in this algorithm are pointed out.(3)This paper proposes a valuate model for mining the weighted association rules algorithm,which based on time clustering. On these bases,we use relationship operation on the boolean vector, and propose an improved algorithm for mining the weighted association rules which based on cluster and compression matrix——CCMW algorithm.The algorithm reduces the size of the database,which based on affairs compression and cluster.It scans the database only once and does not directly generate candidate itemsets to get frequent itemsets.Experimental results show that the efficiency can be improved to over 50% under same working conditions.(4)The design uses CCMW algorithm and Delphi 7.0 on the original management system for supermarket,and design a data mining system for shopping basket analyse of Wan Jia supermarket,which is comprise of data importing,Data Preprocessing,association rules mining,inquiry and analysis.Data importing module can input data to data warehouse which manager is interested in;Data Preprocessing is the process of data cleaning, conversion and data loading in the data warehouse; Association rules mining module can find the association rules that are content with weighted support and weighted confidence,which provides the decisive support for the arrangement of goods. Inquiry and analysis module can inquire some information about goods,and then propose some Marketing Strategies. This system have been put into operation in Wan Jia supermarket,and sales volume increased by 10% a month.