The Design And Implementation Of University Performance Management And Analysis System Based On Decision Tree

Nowadays,most of vocational colleges are still using the traditional educational administration system,the student achievement management uses the simple data statistics,query and backup operation and it can't be used to make deep analysis of student achievement.The aim of this paper is to design a performance management system for our higher vocational college based on decision tree,which not only includes the basic data processing function,but also can use the students' scores to get the relationship between the courses that can help teachers and teaching managers to provide early warning information on student achievement.According to the analysis of the actual situation of the present educational administration system,this paper presents an advanced university performance management and analysis system which has an effective warning function.This system uses the.NET platform and establishes a database for the system.It will focus on the depth analysis of the early-warning system on the students' achievement.The C4.5algorithm of decision tree algorithm is selected,the decision tree model is constructed,and the corresponding rules is extracted.Then according to the students' academic performance,the system provides the corresponding early warning information which offers useful information for the teachers to improve their teaching quality and scientific basis for the university managers to improve the management.This paper uses the C4.5 algorithm in decision tree.Considering its own shortcoming of complicated calculation,the system improves it and adopts the information gain ratio to choose the attributes.In order to reduce the information entropy and information gain ratio calculation complexity,it transforms logarithmic into four arithmetic operations.The sequence of each node is determined only by once calculation and then the decision tree can be generated.In this way,the complex calculations are avoided and the performance of the system data analysis is improved.