The Applied Research Of Customer Relationship Management Based On Association Rules

In this paper, the current commonly used data mining techniques, and their applications in CRM have been analyzed. The typical association rule mining algorithm Apriori also has been deeply studied. A main problem for Apriori algorithm is, it may often create a huge number of itemsets (and rules), and the general users will hard to dig out useful rules. Besides this, it may scan the transaction database repeatedly to cause the heavy operations on I/O. For each k cycle, every element in itemsets Ck must scan the database once to verify whether it join the LK. Referring to the limitation of Apriori, considering the database scanning number, in this paper, a improved Apriori named Fast Apriori algorithm has been proposed. In the Fast Apriori, only once scanning could separate and statistic the entire database in the process of searching for the largest frequent itemsets. When finish the entire database scanning and the counting, system will derive the candidate itemsets from the operational data sets (C1, C2 ... Ck), then, will prune the candidate itemsets according to the minimum support and the confidence in order to identify the frequent itemsets from 1 to k.Compared with the original Apriori algorithm, the Fast Apriori greatly reduce the database scanning number, and need not operate merge and trim. After scanning databases, all the frequent itemsets {L1,L2…Li,(i=1,2,…k)} could be confirmed according to the minimum support and the characteristics of frequent itemsets, and the maximum frequent itemsets could be given through the pruning of frequent itemsets. The strong association rules could be given based on the maximum frequent itemsets and the confidence. The improved algorithm not only could promote the efficiency and quality of data mining, but also could greatly reduce the system resource occupied rate.The applications of association rules in CRM have been analyzed in this paper. The detail realization of Fast Apriori algorithm in CRM customer classification has been analyzed by a case .