A Simple Clustering Knowledge Presentation Method on High-dimension Binary Data Set

摘要

The presentation and explanation of the clustering result play an important role in the technology of clustering . Based on the rough set theory on attribute space, a new clustering result presentation method is advanced. Firstly, the different properties of high-dimension binary data on object space and attribute space have been studied; secondly, the concepts of Low Approximation, Upper Approximation and Feature Precision have been defined for data set on attribute space and the Clustering Information Factor has been defined; and thirdly, a method of Clustering Knowledge Representation on Object Space and attributes space has been proposed. It is simple enough to be understood easily, can provide three kinds of information, that is the distribution of objects, the relationship of the objects distribution and the attributes set, and the rules how to assign new objects to clusters. It can provide relatively synthesis information of clustering result on object space and attribute space, reflect the clustering knowledge with rules, enable users to capture more useful pattern and to hold the internal structure of high-dimension binary data sets.