一种加权的系统聚类方法及应用

A Weighted System Clustering Method and Its Application

提出了加权的系统聚类方法.该方法通过对不同的客户特征赋予不同的权重,达到对客户的聚类结果更符合企业经营目标的目的.为分析客户特征的权重,利用线性回归挖掘企业的历史数据,得到物流企业对客户各个特征的重视程度,并把重视程度作为权重进行客户聚类.实验分析表明,与传统的系统聚类相比,加权的系统聚类方法可以发现隐藏在一般客户中的重要客户,从而使物流企业对本公司的重要客户提供优质服务.

A weighted system clustering method is proposed, which could achieve clustering result for customers more in accordance with the target of an enterprise by given different weights for different customers. For analysis the weights of customer＇s characteristics, linear regression is used in historical data mining of the enterprise to get emphasis of logistics enterprises on the customer＇s individual characteristics, and the emphasis as weights for customer clustering. Experimental results show that compared with the traditional system clustering, the weighted system clustering method can find valued customers hidden in general customers, so that the logistics business could provide high quality services to the company＇s key customers.