Evaluation Method Of Artificial Neural Network-based Education

Education information engineering is an important mission presented for the basaleducation reformation in the period of our country's "the Tenth five-year plane". To useinformation technology in education field help to realize educational modernization andimprove the educational level in our country. Education evaluation is a higherrequirement put forward after the accomplishment of basal information engineeringconstruction in education field. According to some educational value concept andeducation object, using feasible method and systematical collecting information,analyzing and explaining, we can put up value-judgment to education phenomenon.Thereby, it provides warranties for continuously educational improvement anddecision-making.So far from the naissance of education evaluation, according to the effortsresearchers have made, a relatively self-contained system info have been forminggradually。Its evaluation object not only contains early learning effect of students, butalso contains education scheme and activities. Except that, the whole education process isalso contained. Evaluation method contains linear programming, dynamic programming,data envelopment analysis, layer analysis, regression analysis, factor analysis, clusteringanalysis, homogeneous Markov chain and some others.As a main form to check the learning and teaching effect, examination is alwaysadopted by schools. Evaluate the students' learning effect exactly and impersonally byexamination scores is an important part of education evaluation. The objectivity ofexamination not only has to do with the objectivity of student's learning level and abilityevaluation, but also that of student's teaching ability and effect evaluation. But, there aremany impersonality factors that impact on examination scores, such as quality of testpaper, invigilation, mark making, student's ability and so on. Among this, quality of testpaper has directly impact on examination scores. So only scores cannot reflect student'slearning effect exactly.The theory of Artificial Neural Networks is a newly information science developedrapidly in mondial in the medium-term or upper period of 1980s. B-P Network is onestyle of Artificial Neural Networks, which is a multilayer feedforward network, having strong nonlinear mapping ability. Considering the complexity and nonlinear to evaluatethe examination scores synthetically by the evaluation index of the quality of test paper,this paper adopt B-P Neural Networks to modeling and analysis. The primary research isas follows:●Analyzing existing evaluation method of education and emphasizing ondiscussing the methods associated to the questions in Chapter 4; discussing howto use the evaluation methods in synthetically evaluation and decision-makingproblems; analyzing the characteristics and localization of these methods.●Discuss Artificial Neural Networks technology and its application means insynthetically evaluation, there into emphasizing on discussing B-P NeuralNetworks and its arithmetic.●Combined with the student's examination scores synthetically evaluationproblems solving in the Chapter 5, Chapter 4 analyzes the influential factors ofthe quality of test paper and discusses the correlation among the evaluatingindex qualitatively and quantitatively.●Taken the student's examination scores synthetically evaluation problems as theanalyzing object, this paper emphasizes on discussing how to found anintegrated evaluating model that based on B-P Neural Networks and makes asimulation analyzing for the rationality of this model.