The Research Of The Theory Of Extended Rough Sets And It's Application In Risk Assessment Of Industrial Projects

Some scholars have already proposed a lot of dominance relation and its corresponding rough model under incomplete information system in order to deal with the problem of preference decision which has incomplete information.But in the practical application environment the model still have many improvements.This paper try to find the improvement in rough set model based on incomplete information through the analysis of influence factors of misclassification.The new model aimed to improve the accuracy of approximation classification and approximate quality of classification of rough set model.The specific work is mainly reflected in the following aspects:From study we find that the impact of each condition attributes imposed on decision attributes is different.Therefore,the dominance relation based on weight and its rough approximation model is presented.The method of how to set the weight of condition attributes based on rough set and how to set the threshold value have also been put forward in this article.Though the comparison analysis,the weight dominance relation fits the realistic situation better than current expansion dominance relation.The given example has proved the practicality and effectiveness of the weight dominance relation and its rough approximation model.In the incomplete information system,no matter use any dominance relation to construct a rough set model the result is uncertain.In order to reduce the misclassification caused by uncertainty,this paper put forward the concept of gray dominance relation and its rough model.Firstly the analyses of the sectors which cause the uncertainty of classification are presented in this paper.Secondly the concepts of difference coefficient and gray dominance relation are definite.Besides,rough approximation model based on gray dominance relation is constructed.Finally,through the previous example new model proves to have better classification performance.Use the method of how to set the weight of condition attributes and rough set model based on grey dominance relation to analyzing industrial information system.