Mining Frequent Strong Common Families in Weighted Leaf-Labeled Trees

摘要

Tree structures are one of widely used data representation in various domains, which could effectively describe some hierarchical relationship. There have been considerable efforts to identify meaningful patterns from large volume of trees in the data mining literature. This paper is concerned with mining frequent families for weighted leaf-labeled trees in which only leaf nodes have labels and internal nodes have no labels but weight that indicates how strongly the subtree rooted at the node is combined. It presents a mining algorithm proposed to find frequent strong common families from the weighted leaf-labeled trees.