Evaluation Method Consider Expert Information Redundancy

Comprehensive evaluation has many applications in the fields of economics,management,engineering,et al.Comprehensive evaluation needs to introduce a large number of expert knowledge and experience in the evaluation of its complexity determines,at this time,some problems will follow.First of all,index set structuring is the basis of the comprehensive evaluation that could complete scientific and reasonable.It is necessary to deal with the redundancy of attribute information in index set.Secondly,evaluation method of the expert group and the reasonable aggregation of evaluation information affect the objectivity of the comprehensive evaluation directly.Finally,as some attributes are difficult to make accurate judgment because of the uncertainty of environment,the design of fuzzy evaluation method is involved.This paper discusses the above problems and designs the corresponding methods of uncertain multi-criterion group evaluation.It is a common problem that the attribute information redundancy processing of index system in the establishment of index set.For evaluating information that have determined value of the attribute reduction problems,there has been several more mature statistical methods.And for evaluation of uncertain and imprecise environment,the rough set method has a great deal of exploration and application.The above methods have their advantages and disadvantages.The existing method which can handle index value that is not determine or not sure is rough set method which is more mature in the reduction methods.But the description of the index redundancy is on the perspective of set,which is seeking the minimalist index set can cover all attributes.Based on the perspective of single index and the theory of mutual information to measure the index'ability of description target.Being a set of information and according to the law of a class of each index information covered by the other all indicators to measure the ratio of the redundency infoemation,and establish the index selection model,with the goal to minimum redundancy and realize a comprehensive description of evaluation target indicator set at the same time.The index screening method in the article can apply to determine evaluation value and that is uncertain or fuzzy,and the index of redundancy measure could expand existing index screening method's development.In the process of carrying out the evaluation agreement,information redundancy general because of evaluation experts behavior that is unjust produced by taking into account their own interests.And in multi-attribute evaluation,there are always some experts don't have considerable knowledge in some areas,the reference value for the judgment of them are smaller.For the former,this paper use the closed loop multi-round interacting to negotiation and determine the evaluation value to reduce the negative behavior of experts so that ensure the fairness of the evaluation result.In the latter case,this paper discusses two methods of evaluating information aggregation based on the influence of experts: several rounds of interaction determine Expers' influence and network node importance sorting method to determine Expers' influence,and the specific expert influence determination method is given.Among them,expert evaluation values are given in the form of triangular fuzzy numbers and corresponding calculation methods are given.Finally,this paper examines and illustrates the design method and model of this paper with an example analysis,and verifies the feasibility of the model of attribute information redundancy processing and the evaluation information aggregation method bases on expert influence.