A Novel Concise Representation of Frequent Subtrees Based on Density

Abstract

Frequent subtree mining has wide applications in many fields. However the number of the frequent subtree is often too large because of the extensive redundancy in frequent subtree set, which makes it difficult to be used in practice. In this paper, density of frequent subtree in the lattice induced by frequent subtree set is introduced, then a novel concise representation of frequent subtree called FTCB is proposed, and the corresponding mining algorithm FTCBminer is proposed too. Experimental results show that FTCB keeps more information than MFT and reduces the size of frequent subtree set more efficiently than CFT.