Research And Solution Of University Classroom Teaching Quality Evaluation System Based On BPNN

How to improve the teaching quality in higher education, has become the current focus of the work of higher education. However in universities, classroom teaching is the main channel for the implementation of education. Its quality at a large extent reflects and determines the quality of education in colleges and universities. The evaluation is key to improve teaching quality. So how to set up a scientific justice evaluation university classroom teaching quality system is a very important problem.Teaching is a synthesis of dynamic process of teaching and learning, there are many factors affecting it, and they affect it in different degree, the result of evaluation is hardly to use the equal math's analytic expression to show, it belongs to complicated non-linear sort problem, it has brought the very major difficulty for the quality synthetic evaluation. In the past, many the evaluation systems are direct evaluated by setting up a mathematical model, such as the weighted average method, hierarchical analysis process, fuzzy comprehensive evaluation method and so on. The above methods in the assessment process require influencing factors (evaluation index) with a linear relationship. Thus, it is difficult to rule out a variety of stochastic and subjective evaluation results which are easily bring about many distortion and bias.Artificial neural network as a new technique, it is widely used in all kinds of evaluation problems, due to its ability of inherent non-linear processing, adaptive learning and high fault-tolerance. The Back Propagation Neural Network is the one of many types; it is one style of Artificial Neural Networks, which is a multilayer feed forward network, having strong nonlinear mapping ability. Considering the complexity and nonlinear to evaluate the university class teaching quality, this paper adopts BP Neural Networks to modeling and analysis, the primary research is an follows: 1. Analysis of the existing class teaching quality evaluation research, it concentrate on two point, one is on the study of content of quality evaluation system, another is on the Study of the way how to assess quality grade. And has analyzed the characteristics of a variety of evaluation methods and the limitations.2. To introduce the knowledge of neural network, to make a system research on the constructing and training of BP network model. To introduce problems exist in BP learning algorithm and improvements in learning algorithm, Statement corrective method of Levenberg-Marquard BP, and has analyzed its characteristic.3. This dissertation presents a novel classroom teaching evaluation system, which optimizes the quality of design. It includes a comprehensive assessment of teaching quality and in accordance with the main body of the characteristics of universities and disciplines, and it is very useful to optimize the tiered quality of teaching evaluation system.4. Design for university class quality evaluation model. Including fix the network structure, learning parameters and learning algorithm.5. The system carries on the confirmation in simulation software MATLAB, set student evaluation of teaching subsystem as example, realized the functions of style data maintenance, BPNN training and BPNN evaluation. Through the experiment data, it indicates using BPNN evaluate quality is feasible, and satisfy the request of precision, it really prompts a convenience tools for teaching quality evaluation.The paper finally summarized the research work of this article, and discussed some solution of question as well as the further research work prospect.