摘要
以神经网络组合模型和K-means的应用对客户分类和潜在性价值等问题的解决为目的,以CRM理论为前提,对客户潜在价值指标中更具有针对性的指标进行选择,行为习惯和属性两个方面的指标都涵盖在内。完成组合模型的建立,对某个数码网点客户信息及数据进行研究,对于客观属性的分析通过k-means法完成最初的聚类,得到的初类在其内部完成自组织竞争神经网络预测及训练,这样细分聚类就得以完成,然后对聚类结果的特点进行评价,结合具体情况为数码电子商务客户关系管理提出更高效合理的措施与建议。
Aiming to use the combination model of k-means and neural network to solve the problem of potential customer value and customer classification.On the basis of the existing CRM theory,select targeted customer potential value indicators,including objective attribute indicators and behavioral habits indicators.Constructing a combined model to analyze the customer data of a digital online store,using the k-means method to perform preliminary clustering of objective attributes,and conducting self-organizing competitive neural network training and prediction within each primary category to subdivide the clustering results.Finally,the characteristics of the clustering results are evaluated and suggestions are made for digital e-commerce customer relationship management.
作者
高旭
王联国
代永强
GAO Xu;WANG Lianguo;DAI Yongqiang(College of Information Science and Technology,Gansu Agricultural University,Lanzhou Gansu 730070)
出处
《软件》
2021年第2期6-10,共5页
Software
基金
甘肃农业大学研究生重点课程建设和教育研究项目(GSAU-ZDKC-1817,GSAU-ZDKC-1816)。
