| This dissertation, which is finished on the basis of the national fund of natural science ,introduces the theory of soft computing (SC) including ANN(artificial neural network)and fuzzy system (FS) into the analysis of economic contribution of education and analyzes the economic contribution of education from the quantitative view, making use of the ability of highly nonlinear reflection, distributive storage, self-organization, self-learning and self-adaptation of ANN and the characteristics of inferring and making decision like human being ,making full use of human knowledge and effectively dealing with imprecision and instability of FS.Soft-computing(SC), a brand-new method of constructing computing intelligence, is becoming the focus of people. It comprises of ANN, fuzzy set theory, approximate inference, and some non-derivative optimize, such as genetic algorithm and simulation anneal. It can learn from unknown or changing environment and make adjustment improving the result of learning from imprecise and instable environment through coordinating with each composing part. In SC, ANN is responsible for the self-adjustment acting on changing environment and fuzzy inference system makes decision and inference on the basis of human knowledge. This two methods are reciprocity, enabling soft-computing self-adjust and learn better acting like human being in specific fields, moreover, making an explanation how to make decision and take action.The traditional theory of new classical economics(it thinks that the increase of economic is attributed to three factors ,that is, labor, capitals and natural resource)can not explain the reason of economic increase well. The phenomenon puzzled most economists, and at the same time it fastens the development of educational economics and increase economics. The explanation for the phenomenon leads to two sides, the representative of one side is professor Solo, they stands for increase economics, emphasizing the role of technology, learning, human resource for economic increase and modeling and making them endogenetic; the other side is education economics which representative is professor Schultz and professor Denisen,.They emphasize education, human resource. Although their methods are different, the outcome is the same ,that is ,it is important that the role of education for economic increase.Considering the characteristics that the effect of education for economic is unseen, lasting period and multi-sided and the fact of the multi-sided and reciprocal complication, it is a challenge that measuring the education contribution for economic system which is a complicated, large, fuzzy and unstable system. Research of the project has positive effect on theory or practical side because of its successful introduction of advanced mathematics knowledge such as ANN and FIS into social science field.This dissertation analyzes the educational contribution to economic increase from quantitativeand qualitative aspect using economic and computer knowledge and ANN and fuzzy system theory as a tool. Firstly, the dissertation generally discuss the basic notation and characteristics of SC, the history, proficiency and trend of ANN and FIS, then, from quantitative and qualitative aspects, driving role and policy sides , analyzing the educational contribution to economic increase., then, discussing the traditional evaluating method. Then, introducing two model in the project : SOM network and FNN, introducing the construct ,principle, training algorithm and shortcoming of ANN and FNN. The former is used for classification and the latter for building the fuzzy inference system on the basis of former classification. Finally, the dissertation present the characteristics of Matlab language, the theoretical basis of classification and building FNN, then, program using Matlab and make a simple discussion of outcome.The dissertation successfully analyzes the economic contribution of education using SC method from quantitative view, the introduction of these advanced math...
|
PDF Full Text Request