항목중요도를 반영한 가중치 연관규칙 알고리즘

Abstract

The discovery of association rules has been found useful in many applications. The items in a basket database have been examined the same as shown in previous studies. They are different from real world databases because they carry different items each having their own significance. It certain occurances weighted association rules did not reflect the items or their true value. Mining negative association rules have become a focus in the field of data mining recently. The positive association rules could only find in the past weighted association rules, whereas the negative weighted association rules are just as important as the positive ones. we proposed the algorithm that generate frequent itemSets consulted importance of items in this paper. Also, we expand the algorithm in mining for both negative and positive weighted association rules based on correlation. The results of the experiment in simulation show that the frequent itemSets include both for weighted items and unweighted frequent itemSets. Also, the positive and negative association rules increased. The outcome suggests the algorithm is effective.