基于多分类器的房地产客户分类

Classification of Real Estate Client Based on Multiple Class Selectors

利用数据挖掘中分类的技术,根据房地产客户的信息,对客户购买力、购买欲进行评估,根据客户是否会购买,将其分为两类:重点客户和一般客户。从损失成本和辨别能力方面考虑,构建了一个组合分类器模型。使用Weka软件,利用多个公司的客户历史数据,与决策树、神经网络、支持向量机以及贝叶斯网络的分类性能做了比较,发现该组合分类器在稳定性、正确率方面优于其他分类器。

The classification technology is utilized, evaluating the purchasing, power and the desire to buy of client, according to the basic information of client. Clients are classed into important client and general client. A multiple class selectors are constructed, considering the lost cost and resolving ability. Using the data of some companies,Weka is used to compare the performances of this class selector with decision tree, BP, support vector machine and Bayesian network. It＇s found that a high accurate level and stability can be achieved by this class selector.