Incomplete multi-granulation rough set model based on valued tolerance relation

Mengya LIN,Sheng YAO,Ju CHEN,Jie WANG,Feng XU
2017-01-01
Journal of Computer Applications
Abstract:In the incomplete information system,according to the advantages of valued tolerance relation and multigranulation rough set model,a multi-granulation rough set model based on valued tolerance relation was proposed.Firstly,the optimistic and pessimistic multi-granulation rough set models based on valued tolerance relation were defined,and the relative properties were studied.Secondly,the concept of classfication threshold was defined for object classsification;and in different granularities,multiple classification thresholds made the proposed model more flexible.Finally,experimental results of UCI data sets show that compared to the model with a single classification threshold,the multi-granulation rough set model with multiple classfication thresholds has better approximation effect,in each data set,approximation accuracy has the improvement of 0.05 to0.1.
What problem does this paper attempt to address?