文档介绍：长春师范学院
硕士学位论文
基于神经网络的教师教学质量评价模型的研究与实现
姓名:刘军
申请学位级别:硕士
专业:课程与教学论
指导教师:于繁华
2011-05-25
摘要

目前对于高校教师教学评价,还没有一种评价方式是极其公平、合理和科学的。为
使其评价更具有科学性,通常将数学的方法引入高校教师教学质量评价中,一般包括:
传统统计分析评价模型、模糊评价法、灰色决策法、层次分析法(AHP)等等。由于教师
教学评价是一个非线性的问题,上述方法虽然取得了一些好的结果,但是这些方法又都
具有一定的局限性,不论是在评价指标的选取上,还是在评价指标权值的设置上都不够
科学和客观,存在一定的主观性。
人工神经网络凭借其本身具有的非线性处理、自适应学习、高度容错能力等特性在
诸多领域中得到了广泛的使用。BP神经网络是神经网络中的一种,具有很强的非线性映
射能力,因此对于解决教师教学评价这种非线性关系问题具有很强的可行性和科学性。
在建立教师教学评价体系的过程中,本文将主要从不同学科专业考虑,设置不同的评价
指标项目,建立不同的评价指标体系。利用 MATLAB 工具箱强大的功能,建立 BP 神经网
络,进行训练网络、测试网络、最终分析实验结果。
经过实验论证,基于 BP 神经网络的教师教学评价模型在教学评价方面的应用是科
学的、客观的和合理的。

关键字:教师教学评价;人工神经网;BP 神经网络;MATLAB 工具箱

I
Abstract

In present time, there is not a very fair, reasonable and scientific way for the teacher
teaching assessment in colleges and universities. The quality of teacher teaching assessment
in colleges and universities experienced the following methods: Traditional Statistic Analysis
Assessment Model, Fuzzy Evaluation Method, Grey Incidence Decision-making Method, and
Analytic Hierarchy Process (AHP) and so on. Teacher teaching assessment is a nonlinear
question, the methods which mentioned above have received very good result in a certain
time, but these methods all had their own ings. No matter in the setting of
assessment indicators or the weights in the assessment indicators, both are not scientific and
objective, they are both subjective.
Artificial work has applied popularly in all sorts of assessment problems
because it’s strong properties, such as nonlinear process, adaptive learning and a higher
fault-tolerant ability. Back work is a kind of work, which has a
very strong nonlinear mapping ability, so it has very good feasibility and scientific in solving
the nonlinear problem of teacher teaching assessment. In the progress of setting teacher
teaching assessment system, this thesis would set a different assessment