Association rules mining including weak-support modes using novel measures
Jian Hu, Xiang‐Yang Li · 2009
Abstract:-In association rules mining application, some rules can provide a lot of useful knowledge for us, though these rules have the lower Support, called weak-support mode in this paper. However, in existing Support-Confidence framework, the rules with lower Support will be lost. Thus, this paper puts forward a new association rules mining technique, which sets up the lower support threshold value to ensure the weak-support rules to be mined and applies Csupport measure to recognise weak-support mode. Then, a new measure, called as N-confidence, is used to restrict mining size in generation frequent sets, which can strain away the weaksupport rules without correlation. Furthermore, this paper puts forward a new interesting measure to distinguish from the association rules interesting degree. In order to enhance mining efficiency, a novel algorithm, namely FT-Miner, is presented to discover association rules in a forest by using two new data structures, including UFP-Tree and FP-Forest. The experimentation shows that the algorithm not only mines useful and weaksupport rules, but also has better capability than classical association rules mining algorithms. Key-Words:- Data mining; Association rules; Weak-support mode; UFP-Tree; FP-Forest 1