Max-Clique: A Top-Down Graph-Based Approach to Frequent Pattern Mining

Abstract

Frequent pattern mining is a fundamental problem in data mining research. We note that almost all state-of-the art algorithms may not be able to mine very long patterns in a large database with a huge set of frequent patterns. In this paper, we point our research to solve this difficult problem from a different perspective: we focus on mining top-k long maximal frequent patterns because long patterns are in general more interesting ones. Different from traditional level-wise mining or tree-growth strategies, our method works in a top-down manner. We pull large maximal cliques from a pattern graph constructed after some fast initial processing, and directly use such large-sized maximal cliques as promising candidates for long frequent patterns. A separate refinement stage is needed to further transform these candidates into true maximal patterns.