Abstract: For some shortcomings of the hierarchical clustering method and K-Means clustering method, an improved clustering algorithm called DS-Ward is proposed, which based on density biased sampling analysis. The algorithm can automatically generate the center and number of clusters, it can obtain reliable results in the case of reducing the amount of calculation. Through seller’s credit clustering model based on this method, the seller’s sales characteristics of different categories can be found in actual data analysis.

Key words: density biased sampling, partition clustering, hierarchical clustering, credit

摘要： 针对层次聚类方法与K-Means聚类方法的一些不足，提出了一种基于密度偏差抽样的改进聚类分析算法DS-Ward，该算法能够自动获得中心点和聚类数，能够在计算量减少的情况下得到较为可靠的结果。通过基于该方法的卖方信用聚类分析模型对实际数据进行分析，以发现不同类别卖方的销售信用特点。

关键词: 密度偏差抽样, 划分聚类, 层次聚类, 信用