The Generalized Composite Rough Set Model And Its Application

Composite rough set model is a tool for reasoning data which can deal with uncertain,imprecise,incomplete and inconsistent information.Composite rough set model can apply to composite information systems which contains multiple kinds of data.For this reason,composite rough sets is one of the main issues in data mining at the present time.However,on one hand,composite rough set model is lack of adaptability to noise data.On the other hand,composite rough set model can only divide the domain of discourse.In this paper,for the above drawbacks,we proposed three kinds of extended model,i.e..They are variable precision composite rough sets model,multiple variable precision composite rough sets model,composite covering rough sets model,and do a systematic study on every kind of model,which they combine composite rough set model with variable precision rough set model,covering rough set model.The concrete research results are as follows.(1)In view of the problem of the noise data in the composite information system,and the strict requirements of approximate boundary in composite rough sets.Variable Precision Composite Rough Set Model(VPCRS)is presented.The definition of variable precision composite rough set model is proposed by introducing the threshold parameter,and the relevant properties of the variable precision composite rough sets are studied.The application of the model in information processing is illustrated by an example.Further it shows that the model has a certain fault tolerance and enhances the anti-jamming capability,which enlarges the range of application.(2)Due to that variable precision composite rough sets model is proposed by introducing the same threshold parameter for all sets of attributes in the composite information system.Which leads to the model lack of flexibility.In this paper,we propose Multiple Variable Precision Composite Rough Sets(MVPCRS)by introducing the different threshold parameters?_i(0??_i(27)1),(i(28)1,2,...)for different sets of attributes in the composite information system.Then we combine the results of intersection operation of the corresponding variable precision rough set operation for each type of attribute set.Then the related concepts and properties are presented.Finally,the application of the model in information processing are illustrated by an example.Further,the results show that the flexibility of the model is enhanced.(3)Because composite rough set model can only divide the domain of discourse and the classical covering rough sets can only apply to single data type in the covering domain.Composite covering rough sets are presented by establishing a variety of coverage relations(neighborhood covering relation,set-valued covering relation,the interval-valued covering relation,etc.)under the composite data model.The relevant concepts and properties are presented.Finally,the application of the model in information processing are illustrated by an example.Further,the results show that the model conforms to the actual situation and is practical.