摘要
采用一种改进后的决策树归纳聚类算法和交互式CLTree(ClusteringbasedondecisionTrees)剪枝,对商业数据的某些问题实现了聚类挖掘。对交易数据的实际聚类分析表明,该方法不仅可以处理数值型属性,还可以处理枚举型属性。实验结果表明,该方法在处理混合类型数据时具有良好的挖掘效果。对商业数据聚类分析,可以得到合理的市场分段,预测顾客的购买行为。
With a improved algorithm using inductive decision tree and techn ol ogy of alternating CLTree(Clustering based on decision Trees) pruning, the clust ering data-mining is carried out for some aspects of the business data. The numerical type and the enumerating type can be processed. The experiment result indicates that it has good mining effect in processing the mixing type data.The reasonable market subsection and will be gotten and th e purchasing behavior of customers will be forecasted in the business data analy sis.
出处
《吉林大学学报（信息科学版）》
CAS
2003年第2期132-137,共6页
Journal of Jilin University（Information Science Edition）
基金
国家自然科学基金资助项目(60175024)
教育部科学技术重点项目(02090)
关键词
数据挖掘
决策树归纳
聚类分析
相似度
Data mining
Induction of decision tree
Clustering analysi s
Similar degree
