A Kind of Unusual Customers Recognition System Based on Multi-criteria Neural Network and CART in Telecom System
基于多准则神经网络与分类回归树的电信行业异动客户识别系统

As the telecom system an application object, this paper presents a kind of unusual customers recognition system based on multi\|criteria neural networks (MCNN) and classification and regression tree (CART) in telecommunication system. The system reduces customers' attribute by multi\|criteria neural networks first. Then it constructs classification and regression trees to recognize unusual customers. Through actual validating to large customer data in Zhejiang telecom system, the results indicate that this system is more robustious and more effective.