Application Research With Dea In Model Evaluation

The method of DEA has been widely used in various fields for social production and social life, not only includes the relative management efficiency and comprehensive evaluation, economic system modeling and parameter estimation, the profit on the cost-benefit analysis, forecasting and early warning , system of classification and control, but also includes the evaluation of the progress of technological innovation for technology and productivity,the relative effectiveness of the different programs and so on. For the same question we tend to use different methods to establish different mathematical models, so we need to evaluate these models and compare the relative effectiveness of these models in order to decide which of the model for this issue can get better results. Based on the DEA method, method of trying to use DEA to compare the same issue of the relative effectiveness of different models were studied.In the first chapter,this article first introduces the practical background of the problem, and give a brief introduction of the development of DEA methods and the status quo of research.In the second chapter, we detail the basic model of DEA methods - C~2R model and C~2GS~2 model, and give the general steps for the application of DEA method, finally in accordance with the characteristics of DEA we state that the application of the DEA approach to evaluate the "merits and demerits" of different models for the same problem.In the third chapter, we made an empirical analysis on an applied researc of DEA method in models evaluation. We firstly set up three venture capital optimization model of five investment of specific stocks ,and calculated the outcome of each optimization model of venture capital ,using the package Lingo9.0; Then we have the three above-mentioned optimization models of venture capital as three DMUs of DEA method, and consider the extended solver steps and iteration steps based on the Lingo corresponding procedures of each optimization model of venture capital as the DEA input, the average level of expectations and risk level as the DEA output; Finally, we conducted relative effectiveness analysis of modelC~2R, model C~2R of non-Archimedean infinitesimal parameters, model C~2GS~2 , model C~2GS~2 with a non-Archimedean infinitesimal parameters on the three DMU , and obtained the conclusions consistent with experience, that is, model 1 is relatively effective, while model 2 and model 3 is relatively ineffective. Therefore, the use of DEA method to evaluate the model is feasible.