改进的K-means算法及其在铁路客户细分中的应用

Improved K-means Algorithm and its application in railway customer segmentation

提出使用距离均和识别孤立点,并引入方差对孤立点进行判断处理,对传统的K-means算法进行改进并将改进后的K-means算法应用到铁路客户细分领域,实验结果表明,改进后的K-means算法能更为准确地对铁路货运客户进行聚类分析,从多维的角度较为全面、深入地细分客户消费行为特征,从而辅助铁路货运营销部门制定有针对性的营销策略,进行高效的客户关系管理,提高市场竞争力。

This paper put forward the idea of improving the traditional K-means Algorithm with identifying the isolated points by average distance sum, and judged and processed the isolated points by introducing variance. The improved K-means Algorithm was applied to the field of railway freight customer segmentation. The experimental result showed that the improved K-means Algorithm could carry out cluster analysis on railway freight customers more accurately, sectionalize the characteristics of customers＇ consuming behaviors in a comprehensive and in-depth manner from multidimensional perspectives, thus assist railway freight marketing department to formulate targeted marketing strategies, carry out high efficient customer relation management and increase market competitiveness.