Mining Association Rule by Using Prefix Graph with an Efficient Flow-Based Strategy

Abstract

PG Miner, a graph-based Algorithm for mining frequent closed item sets consists of construction a prefix graph structure and decomposing the database to variable length bit vectors, which are assigned to nodes of the graph. The main advantage of this representation is that the bit vectors at each node are relatively shorter than others existing methods. Use projected databases to prune their non-closed item sets. There are two typical strategies adopted by these algorithms: (1) an effective pruning strategy to reduce the combinational search space of candidate item sets and (2) a compresses data representation to facilitate in-core processing of the item sets.