摘要
提出一种使用互连性度量聚类间相似度的层次聚类算法 ,并对算法中较为耗时的两步进行了修改 ,在不牺牲质量的前提下 ,提高了算法的运行速度 .通过分析交易数据的实际聚类 ,可得到合理的市场分段 ,预测顾客购买行为 .实验结果表明 ,该方法具有良好的挖掘效果 .
A clustering analysis algorithm in transactional data-set to use interconnecting cluster is presented. And it is modified for two step of more time consuming. In precondition of ensuring (quality), the running rate is improved. So, it gains reasonable market subsection and forecasts purchasing of customers, by clustering analysis to transactional data-set practically. Experimental result making clear, the algorithm has nicer effect clustering.
出处
《深圳大学学报（理工版）》
EI
CAS
2003年第1期63-69,共7页
Journal of Shenzhen University(Science and Engineering)
基金
深圳大学科研基金资助项目 (2 0 0 1 2 8)
