摘要
银行客户群体细分对于业务营销具有深远的意义,客户信息具有数据量大、维度高、变化需求频繁的特点,为此需要引入一种快速的属性约简算法,以满足关键属性快速提取进而构建决策的要求.本文通过改进传统的基于蚁群的属性约简算法,优化每次迭代过程中的蚂蚁搜索的集合转移策略,提出了一种基于快速蚁群算法的属性约简算法.多个UCI数据集实验计算表明提出的新算法求解速度优于传统的基于蚁群算法的属性约简算法,并且求解质量较优;最后通过银行客户数据进行实践,验证了该算法的可行性.
As bank customer segmentation has a profound significance for business marketing, while customer information has the characteristics of large amounts of data high dimensions and frequently-changing demand, we need to introduce a fast algorithm for attribute reduction to meet the needs of rapid attribute extraction to construct decisions. This paper proposes a new quick attribute reduction based on ant colony optimization by improving the collection for each iteration of the ant search transfer strategy. Numerical experiments on a number of UCI datasets show that the proposed new algorithm has a lower computational cost than the traditional ant colony-based attribute reduction algorithm and a better solution quality. Finally, the feasibility of the proposed algorithm is verified through the use of the bank customer data.
出处
《计算机系统应用》
2015年第10期217-221,共5页
Computer Systems & Applications
关键词
客户信息
粗糙集
属性约简
蚁群算法
快速提取
bank customer information
rough set
attribute reduction
ant colony algorithm
rapid attribute extraction
