层次分析与聚类分析、 判别分析在卷烟零售户分类中的应用

为了优化卷烟零售户资源配置， 调研了石河子市烟草公司， 构建了卷烟零售户价值评价指标体系， 运用层次分析法确定各指标的权重， 用聚类方法对701户卷烟零售户评价指标数据进行聚类分析， 用判别分析检验聚类结果， 并针对每类客户群的特点提出了相应的营销策略。 结果表明， 701户卷烟零售户可分为9类， 判别分析与聚类结果一致率为95.3%， K中心聚类分析得出的9类客户群是比较合理的， Fisher判别得出的判别函数能有效识别新零售户所属的客户群类别。 零售户价值评价指标体系能有效地表征卷烟零售户的价值， 层次分析结合聚类分析、 判别分析可用于卷烟零售户分类。 

Application of AHP and Clustering,Discriminant Analysis in Tobacco Retailers Segmentation

In order to optimize resources allocation for tobacco companies, according to field survey of Shihezi city tobacco company, the fire-new evaluation system of tobacco retailers was built, AHP (analytic hierarchy process) was applied to determine the weight of each index, clustering method was employed to cluster analysis for the 701 retailers, discriminant analysis was used to detect clustering result, and specific marketing suggestions was put forward for each category of retailers by their marketing characteristics. The research result show that 701 tobacco retailers was segmentalized into nine categories, the consistent rate of discriminant and clustering analysis for 95.3%, the result of K-means cluster analysis is certain reasonable, and the discriminant function obtained by Fisher discriminant can effectively identify the new retailer’s category. In addition the evaluation system can effectively reflect the value of tobacco retailers, and the combination of AHP and cluster analysis, discriminant analysis can be used for tobacco retailers’ segmentation.