Skyline quantity-utility sequential pattern mining: An efficient and effective approach

doi:10.1016/j.knosys.2025.114185

Skyline quantity-utility sequential pattern mining: An efficient and effective approach

2025-08-05

0

PRE

AI

DOI：10.1016/j.knosys.2025.114185

Original

Original request for help

Save

Abstract

En 中文

• We propose SQUMiner, a skyline-based sequential pattern mining algorithm that integrates both quantity and utility for robust decision-making. • SQUMiner introduces novel data structures, QU-array and QUPro, enabling efficient management of sequence quantity and utility values. • The Maximum Utility Quantity Array (MUQA) and Sequence Pareto Front Array (SPFA) are novel physical structures designed to support pruning strategies and optimize the mining of high-quality patterns. • Three pruning strategies (SQUD, IQUD, EQUD) are designed to optimize the mining process by dynamically pruning dominated sequences based on utility upper bounds. • Extensive experimental results show that SQUMiner achieves up to two orders of magnitude speed-up and reduces memory usage by up to 85 % against six state-of-the-art baselines.

Keywords:

SQUMiner

sequential pattern mining

utility mining

pruning strategies

quantity and utility integration

Skyline quantity-utility sequential pattern mining: An efficient and effective approach

Abstract

Journal

Organization