Xu Jiucheng,Shen Junyi,An Qiusheng,Li Naiqian
(School of Electronics and Information Engineering, Xi'an Jiaotong University, Xi'an 710049, China)
Abstract: Based on information granularity and rough set theory, the
relations between the decision subdivision degree and the information granularity, the
accuracy of approximation classification, and the quality of approximation classification
were mainly discussed. It is theoretically demonstrated that for any condition attribute
set, the finer the decision attribute value of a decision table is, the lower the
information granularity, the accuracy of approximation classification, and the quality of
approximation classification are. Simulation results show that the information
granularities, the accuracy of approximation classification and the quality of
approximation classification in the finer decision table are not bigger than the ones
corresponding to the decision table before decision refinement. The research is helpful
for the attribute reduction and for enhancing confidences of decision rules.
Keywords: decision subdivision;information granularity;accuracy of approximation
classification;quality of approximation classification