Association Classification Based on Compactness of Rules
Qiang Niu, Shixiong Xia, Lei Zhang · 2009
Associative classification has high classification accuracy and strong flexibility. However, it still suffers from overfitting since the classification rules satisfied both minimum support and minimum confidence are returned as strong association rules back to the classifier. In this paper, we propose a new association classification method based on compactness of rules, it extends Apriori Algorithm?which considers the interestingness, importance, overlapping relationships among rules. At last, experimental results shows that the algorithm has better classification accuracy in comparison with CBA and CMAR are highly comprehensible and scalable.