Improved Algorithm of Multiple Minimum Support Association Rules Based on Can Tree

Abstract

To improve the mining efficiency of Can Tree in association rules, this study found that the header table in the traditional Can Tree mining algorithm repeatedly traverses the path, causing excessive overhead. And the single support will lead to invalid rules. The improved Can Tree mining algorithm performs reduction processing and removes the header table, making it more suitable for the idea of Can tree construction. Traversing the path only once avoids multiple traversals, and the addition of multiple minimum support does not introduce too much overhead. Experiments have proved that the improved Can tree mining is better than the traditional mining algorithm.