Comprehensive Evaluation Model System Construction And Its Empirical Study

Practical problems, a thing often requires multiple indicators to characterize its essential character, according to multiple indicators attribute them to make a comprehensive assessment and decision-making, the results of the use of vague language is divided into different levels of review, which is a comprehensive evaluation. Comprehensive evaluation model system should include the selection of the evaluation, a comprehensive evaluation index weights determined, as well as evaluating the object of three parts.By Delphi method to establish the hierarchy of the evaluation, taking into account the fuzzy data and subjective, assuming that the choice of experts on each indicator is a fuzzy interval value a=[a-,a+], this choice of alternative indicators constitutes a fuzzy matrix M for elements a. Finally, adjust the sensitivity of the classification level λ=[λ-,λ+] and class h0, the alternative indicators are divided into classes. Its essence as the domain of alternative indicators, choose the indicators will comment on some standard domain elements are divided into classes, each class indicators reflect the mix of attributes. Determination of the classification level λ means the index system level determination. Generic sensitivity h0means that the interlayer fine-tuning of the index clustering. This paper proposes a similarity coefficient rij of interval-valued scalar product method to construct.Evaluation index weight determined fuzziness between the data and the assumptions of experts on each of the indicators of the right to re-determine a fuzzy interval value. Thus, the index weights constitute a range value of fuzzy matrix elements. Fuzzy set-valued statistical method is used to calculate interval value frequency for the coverage, and the relative weight W of the index is obtained. Finally, the number of consolidated indicator as generic weight, by variable weight function W=f(N,W), structural composite weights of evaluation indexes. Ambiguity or reliability is important yardstick of the relative weights of whether to meet the requirements.Judge of the evaluation object, the more mature methods is fuzzy evaluation, due to the computing attribute fuzzy operator, resulting in the loss of a lot of useful information. On the other hand, linear estimation functions have also led to some cases, the evaluation results failure. Set on the number of links in the analysis is the number of a system is layered, fuzzy, reflecting the state and evolution of trends. This paper argues that the vagueness and uncertainty of human thinking outset describe with certainty the amount of rough fuzzy evaluation results. Therefore, the definition of fuzzy evaluation of the credibility of the concept as well as with the credibility of the number of links, using fuzzy evaluation and contact number of momentum principle, to establish a static evaluation model, the static sorting model as well as the dynamic evolution model.Finally, specific issues as the research object, this model system is an empirical. The empirical results show that the system of comprehensive evaluation model has a certain operability and practicality. Evaluators of the same class, the evaluation of the behavior of objects having a certain predictability.