A Tight Upper Bound on the Number of Candidate Patterns

TL;DR

This paper derives a tight upper bound on the maximum number of candidate patterns generated in levelwise frequent pattern mining, aiding in reducing database scans and improving efficiency.

Contribution

It introduces a novel tight upper bound based on classical combinatorial results, enhancing the understanding of candidate pattern generation.

Findings

Provides a mathematically proven upper bound for candidate patterns

Helps optimize the pattern mining process by reducing unnecessary database scans

Connects combinatorial theory with practical data mining algorithms

Abstract

In the context of mining for frequent patterns using the standard levelwise algorithm, the following question arises: given the current level and the current set of frequent patterns, what is the maximal number of candidate patterns that can be generated on the next level? We answer this question by providing a tight upper bound, derived from a combinatorial result from the sixties by Kruskal and Katona. Our result is useful to reduce the number of database scans.