The Application of Genetic Algorithms and Multidimensional Distinguishing Model in Forecasting and Evaluating Credits for Mobile Clients

To solve the arrearage problem that puzzled most of the mobile corporations, we propose an approach to forecast and evaluate the credits for mobile clients, devising a method that is of the coalescence of genetic algorithm and multidimensional distinguishing model. In the end of this paper, a result of a testing application in Zhuhai Branch, GMCC was provided. The precision of the forecasting and evaluation of the client’s credit is near 90%. This study is very significant to the mobile communication corporation at all levels. The popularization of the techniques and the result would produce great benefits of both society and economy.

基于客户信用评价模型的营销管理系统

提高企业营销管理的质量和效率,对客户信用评价模型和企业营销管理系统设计方法进行了研究。提出了一个基于马尔可夫逻辑网的客户信用评价模型CCEM-MLNs,该模型将依据专家知识建立的信用评价规则转化为一阶逻辑规则,构造马尔可夫逻辑网,然后利用样本数据进行训练,得出各规则的权值,建立评价模型。实验结果表明,该模型对客户信用分类的准确率比均分权值的评价模型提高了13.38%;设计并实现了一个基于CCEM-MLNs的营销管理系统MMS-CCEM,给出了系统的体系结构和层次结构,该系统已成功应用于国内某企业的营销管理,在客户关系管理方面效果明显。