Abstract: According to the analysis and prediction of the trend of support vector and confidence vector in the mining of dynamic association rules, an improved Grey Model of Particle Swarm Optimization（PSOGM） with buffer operator is proposed. Due to the introduction of the background value, which leads to the declining accuracy of the gray model prediction as well as some limitations, it is necessary to introduce the parameters to modify, by using the improved particle swarm optimization algorithm which joins two search mechanism to improve the local search ability of the algorithm, and background values of the gray model are optimized at different times, then the prediction accuracy of grey model is improved. Through the experimental simulation on the Matlab platform with the data set, the results manifest that this method is more accurate than the original grey model, grey model with genetic algorithm and the standard grey model with the particle swarm optimization.

Key words: particle swarm optimization, gray model, dynamic association rule, background value

摘要： 针对动态关联规则挖掘中支持度向量和置信度向量变化趋势的分析和预测，提出一种改进的粒子群优化的灰色模型应用在动态关联规则挖掘中。由于灰色模型在引入背景值后导致在非平稳序列中的预测精度下降，因此有必要引入参数进行修正，通过在粒子群优化算法中引入二次搜索机制，优化求解灰色模型不同时刻的背景值，从而提高粒子群算法的局部搜索能力，进而提高灰色模型的预测精度。通过在Matlab平台上进行实验仿真，数据集采用超市购物数据，结果表明该方法比原始灰色模型、遗传算法优化的灰色模型和标准的粒子群优化的灰色模型具有更高的预测精度。

关键词: 粒子群优化算法, 灰色模型, 动态关联规则, 背景值