An United Extended Rough Set Model Based on Developed Set Pair Analysis Method

摘要

Different with traditional set pair analysis method, a new method to micro-decompound the discrepancy degree is proposed in line with the distributing actuality of missing attributes. Integrated with rough set theory, advances the united set pair tolerance relation and gives corresponding extended rough set model. Different values of identity degree and discrepancy degree can modulate the performance of this model and extend it's application range. Then expound some existed extended rough set model are the subschema of it. At last simulation experiments and conclusion are given, which validate the united set pair tolerance relation model can improve classification capability.