An improved approach for weighted association rule mining based on multi-criteriaimportance

Association Rule Mining (ARM) is used to identify relationships amongst items in transactional databases so it is widely applied to the subjects which are difficult to use statistical and mathematical methods in recent years. A traditional association rule mining (ARM) are assumed to have the same importance for each item included in the transactional database. In fact, the data in the reality vary with the importance of each data and also it is duo to how the criteria and methods of evaluating them are set. Thus, taking this into account, the mining of association rules can lead to more efficient and accurate results. From this point of view, a number of studies have been carried out to determine the Frequent Weighted Itemsets (FWI) by assigning the weight to each item in the transactional database and to mine the association rules. In this paper, we propose a method for determining the weights based on the multi-criteria importance of items in a transactional database. This paper also proposes the method for weighted associated rule mining (WARM) which is based on the multi-criteria importance weight and some applications are introduced to verify its effectiveness. The experimental results show that the proposed method is better than the typical associated rule mining or previous WARM in accuracy and effectiveness. Therefore, we can be sure that the proposed method is more accurate and effective than the previous methods.