【摘要】 大量布局运力来满足客户需求的模式,使得配送资源无法得到合理利用,针对此问题提出对城市快递客户进行聚类分析。面向客户的消费能力和地理位置,确定了客户聚类的6个指标。基于谱聚类算法,结合经验规则k≤n^1/2,计算Laplace矩阵最大特征值差,确定客户聚类数目k。运用k-means算法对前k个最小特征值对应的特征向量聚类,确定具体的聚类方案。最后,以武汉市某区的36个街道为例验证了算法的有效性,为城市快递客户聚类分析提供理论依据。更多还原

【Abstract】 Aiming at the problem that the operation mode that meets the customer demand through a large number of layout capacity makes it impossible to make rational use of the distribution resources,a clustering analysis of the urban express customer is presented. Considering the consumption ability and geographical location of the customer,six indexes of customer clustering are determined. Based on spectral clustering algorithm,combined with the empirical rule,the customer clustering number is determined by the maximum eigenvalue difference of Laplace matrix. The K-means algorithm is used to cluster the feature vectors of the first minimum eigenvalues,and the specific clustering scheme is determined. Finally,taking 36 streets in one district of Wuhan as an example,the effectiveness of the algorithm is verified,which provides a theoretical basis for urban express customer clustering.更多还原