An Application Research Of Teaching Quality Evaluation Classification Based On SVM

Imparting knowledge and educating people is the duty and responsibility of ordinary institutions of higher learning, and the teaching quality is their lifeline, which is the focus people concerned most about higher education. But, with the expanding of college enrollment, the contradiction between enrollment and teaching quality is more significant. So, new requirements for the teaching quality come out, this is not only need to control the teaching process, but also to establish a set of comprehensive teaching quality evaluation system. Due to the development of information technology(IT), teaching management systems have been widely used at colleges and universities, which continues to accumulate so much data about teaching. How to make use of these data to improve the quality of teaching becomes a question for instruction administrators. This paper considered the practical and economic factor, started from he author’s real work content, and combined with data mining(DM) technology and support vector machine(SVM) theory, aims to establish a suitable classification method for teaching quality evaluation of higher school.(1) The current situation of college teaching quality valuation problem has been studied, the main problems existed have been put out, and the demand analysis has carried out. Considered with the real situation of Zhejiang Business College, an evaluation index for teaching valuation problem has been established in this paper. The classification of teaching quality evaluation is a multi-classification problem, classification analysis method to classify the evaluation data referred to the basic principles and concepts of data mining was used, and several major classification methods were analyzed. (2) The concept of support vector machine was introduced for teaching quality valuation problem. Take advantage of partial binary tree multi-classification fewer classifier and short train time, a partial binary tree SVM method was designed for data classification. Compared with other multi-classification methods, the experimental result shows that the algorithm proposed in this paper has advantages inaccuracy and training time.(3) For the problem studied in this paper, a teaching quality valuation system base B/S was designed and implemented. The system developed referred to the software engineering process, with the ease operation, maintainability, and scalability.