Discretization Algorithms of Rough Sets Using Clustering
Chengdong Wu, Mengxin Li, Zhonghua Han, Ying Zhang, Yong Gang Yue · 2005
In this paper, hierarchical clustering method is introduced for attribute discretization. It can determine automatically the significant clusters. First the best classes for discretization are picked from scatter plots of several statistics. Moreover these classes keep consistent with extracted clusters from dendrograms. By comparison, hierarchical clustering discretiztion method is typically more effective and advisable among several cluster algorithms with the defect inspection of wood veneer.