Mining vague association rules

In many online shopping applications, traditional Association Rule (AR) mining has limitations as it only deals with the items that are sold but ignores the items that are almost sold. For example, those items that are put into the basket but not checked out. We say that those almost sold items carry hesitation information since customers are hesitating to buy them. The hesitation information of items is valuable knowledge for the design of good selling strategies. We apply vague set theory in the context of AR mining as to incorporate the hesitation information into the ARs. We define the concepts of attractiveness and hesitation of an item, which represent the overall information of a customer's intent on an item. Based on these two concepts, we propose the notion of Vague Association Rules (VARs) and devise an efficient algorithm to mine the VARs. Our experiments show that our algorithm is efficient and the VARs capture more specific and richer information than traditional ARs.