摘要： 针对约束频繁模式挖掘中模式指标的界值估算问题,提出一种基于修正二进制粒子群优化(BPSO)算法的通用模式指标上界估算方法。根据带有权值的不确定型事务数据库的特点,建立通用的模式指标上界估算框架,并提出在该框架下基于修正BPSO的模式指标上界值求解方法。对比UHUI-Apriori算法分别结合事务加权效用值、本文方法估算所得上界值和实际上界值后的候选项集数量、运行时间和内存占用情况,结果表明,该方法可以较快计算模式效用的上界值,且能够节省运行时间和内存空间。

关键词: 不确定型数据库, 模式指标, 界值估算, 粒子群优化算法, 约束频繁模式挖掘

Abstract: A general pattern index upper bound estimation method based on modified Binary Particle Swarm Optimization (BPSO) algorithm is proposed to estimate the boundary value estimation problem in constrained frequent pattern mining.According to the characteristics of the uncertain transaction database with weights,a general pattern index upper bound estimation framework is established,and the method of solving the upper bound value of the pattern index based on the modified BPSO is proposed.The number of candidate itemses,runtime and memory usage of the UHUI-Apriori with Transaction Weighted Utilization(TWU),the proposed method estimates upper bound and the actual upper bound are compared,and the results show that the proposed method can quickly calculate the upper bound value of the mode utility and save the running time and memory space.

Key words: uncertain database, pattern index, bound value estimation, Particle Swarm Optimization(PSO) algorithm, constrained frequent pattern mining

中图分类号: 

• TP391