基于支持向量机的高校课堂教学质量评价研究

发布时间：2018-11-25 16:55

【摘要】：高校课堂教学目前是各大高校教学的主要形式,它是高校教学的基础并且在教学过程中具有非常重要的作用。而课堂教学质量评价体系的建立和实施不仅对高校的教学发展理论有很大的帮助作用,更能保障高校课堂教学质量评价的顺利进行以及课堂教学活动发挥有效的作用。目前,传统的教学质量评价方式在只有学生参与的情况下,虽然已经取得了一定的成果,但是还有一些问题没有很好的解决,比如学生的主观因素对教师存在一定的偏见,使得评价结果出现误差,或者只注重评价结果而不能体现教师教学的过程,也会导致评价结果出现误差等等。而支持向量机(Support Vector Machines,简称SVMs)被引入教学质量评价之后,学生、同行和领导都参与评价,不仅能够避免人为因素对结果造成的误差还能充分体现教师的教学过程。另外教学质量评价是一种多类分类问题,最终选择支持向量机多类分类算法对本文的课堂教学质量评价结果进行预测。概括起来,本文的主要工作如下:(1)分析和总结了课堂教学质量评价的意义和传统的评价方法存在的缺陷。根据具体的需求和评价指标体系的构建原则,制定了课堂教学质量的评价指标体系。由于各指标之间存在非线性关系,因此,决定将支持向量机算法应用于课堂教学质量评价中,用来解决教学质量评价中可能遇到的问题。(2)介绍了目前常用的几种支持向量机多类分类算法,重点研究了二叉树支持向量机多类分类算法,并且针对已经存在算法生成的是偏二叉树的缺陷,提出了一种新的改进思想。改进算法利用完全二叉树的生成策略以及聚类中的类距离的相关定义,使得生成的二叉树结构达到完全或者近似完全的状态,从而提高分类精度和效率。最后通过在UCI数据集上做仿真实验,验证了改进算法的有效性。(3)利用改进的二叉树支持向量机多类分类算法,构建基于二叉树支持向量机的高校课堂教学质量评价模型,针对山东省某高校的教学质量进行评价,填写评价量表,并且统计、收集多组数据。在MATLAB环境下,对收集到的数据集进行实验,并分析其结果。将改进算法的预测精度和效率与支持向量机算法、二叉树支持向量机算法相比较,改进算法优势明显,能够更好的预测未标记样本。
[Abstract]:At present, classroom teaching in colleges and universities is the main form of teaching in colleges and universities. It is the foundation of teaching in colleges and universities and plays a very important role in the teaching process. The establishment and implementation of the evaluation system of classroom teaching quality can not only help the theory of teaching development in colleges and universities, but also ensure the smooth progress of evaluation of classroom teaching quality and the effective role of classroom teaching activities. At present, although the traditional teaching quality evaluation method has only the participation of students, although it has achieved certain results, there are still some problems that have not been solved very well. For example, the subjective factors of students have certain prejudice against teachers. It makes the evaluation result error, or only pays attention to the evaluation result but not the teacher teaching process, also will cause the evaluation result to appear the error and so on. After the introduction of support vector machine (SVMs) to the evaluation of teaching quality, students, peers and leaders all participate in the evaluation, which can not only avoid the errors caused by human factors, but also fully reflect the teaching process of teachers. In addition, the evaluation of teaching quality is a kind of multi-class classification problem. Finally, support vector machine multi-class classification algorithm is chosen to predict the evaluation results of classroom teaching quality in this paper. To sum up, the main work of this paper is as follows: (1) the significance of classroom teaching quality evaluation and the shortcomings of traditional evaluation methods are analyzed and summarized. The evaluation index system of classroom teaching quality is established according to the concrete needs and the construction principle of evaluation index system. Because of the nonlinear relationship among the indicators, it is decided to apply the support vector machine (SVM) algorithm to the evaluation of classroom teaching quality. It is used to solve the problems that may be encountered in the evaluation of teaching quality. (2) several commonly used SVM classification algorithms are introduced, and the binary tree support vector machine multi-class classification algorithm is studied emphatically. In order to solve the problem of partial binary tree generation, a new improved idea is proposed. The improved algorithm makes use of the generation strategy of complete binary tree and the relative definition of class distance in clustering to make the structure of the generated binary tree complete or nearly complete so as to improve the classification accuracy and efficiency. Finally, the effectiveness of the improved algorithm is verified by the simulation experiment on the UCI dataset. (3) using the improved binary tree support vector machine multi-class classification algorithm, the evaluation model of college classroom teaching quality based on binary tree support vector machine is constructed. This paper evaluates the teaching quality of a university in Shandong province, fills out the evaluation scale, and collects many groups of data. In MATLAB environment, the collected data sets are tested and the results are analyzed. Comparing the prediction accuracy and efficiency of the improved algorithm with the support vector machine algorithm and binary tree support vector machine algorithm, the improved algorithm has obvious advantages and can better predict unlabeled samples.
【学位授予单位】：重庆师范大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：G642.4

【相似文献】