Mining closed frequent itemset based on FP-Tree

Abstract

Closed frequent itemsets provide a minimal representation of the itemset without losing their support information and they can greatly reduce the number of patterns. So how to obtain all the closed frequent itemsets effectively is of good importance. This paper constructs a FP-Tree according the matrix which only scans the database once. Also we design an algorithm to mine closed frequent itemsets which use the structure of the FP-Tree. We construct a link for every single node in the FP-Tree, and then give the closed frequent itemsets and the pruning of the link. This algorithm is an algorithm which only scans the database once, also it avoids the defect about patterns tree based on terms and can mine the closed frequent items efficiently.