The Applied Research Of Ordinal Decision Tree In The College Students Comprehensive Quality Evaluation

With the continuous development of education and the expending of the education scale of colleges, how to improve the comprehensive quality of students and how to scientifically evaluate the comprehensive quality of students have become the hot concerns by university administrators. The scientific evaluation for Comprehensive Quality of College Students is not only related with the development of the students, but also related with the training the high-quality personnel in Universities.After the analysis and research of the students' comprehensive Quality Evaluation scores, this paper found that the mining of a reasonable assessment rule in essence, is to solve a kind of ordinal classification problem in the field of machine learning. Due to the ordinal relationship in the course scores of students, the results of the Comprehensive Quality Evaluation can be expressed by "fail", "pass", "good", "excellent", and the ordinal relationship in the results of the Comprehensive Quality Evaluation because of that the relationship meet "fail <pass <good <excellent", we believe that there is a monotone constraint relationship between student courses scores and the results of the comprehensive quality evaluation: the better of all the student courses scores, the higher of the grade of the comprehensive quality evaluation results.This paper mainly completed the following work:1, This paper apply the ordinal decision tree based on the ordinal entropy to the students' comprehensive quality evaluation experimental data, and the results show that the ordinal decision tree algorithm has better classification results than the classical C4.5 decision tree algorithm.2, As a routine work, the work of the college students' comprehensive quality evaluation is carried out by the university every year. In recent years, the expanding scale of the colleges and the increasing number of the students lead to the increasing workload of the Students' comprehensive quality evaluation; teaching reform carried out in universities will lead to changes in curriculum, and the curriculum changes will lead to the rediscovery of the college students' comprehensive evaluation rules. To solve the above problems, this paper apply the attribute selection algorithm based on the ordinal mutual information and the maximum correlation minimum redundancy information theory to the Students' comprehensive quality evaluation work for the curriculum selecting, and then use the ordinal decision tree for mining rules; Experimental results show that the rules mined is more stable, the amount of computing for mining rules is greatly reduced, and the efficiency of the evaluation work is improved.3, this paper design to achieve the Students' comprehensive quality evaluation system, and the ordinal decision tree algorithm and the attribute selection algorithm are applied to the function module in students' comprehensive quality evaluation system, which promotes the quality improvement of the students' comprehensive quality evaluation work.