Research On The Decision Tree Algorithm Applied In The Student Evaluation System

With the development of the information technology, the ways to obtain data becomemore and more diverse, and then the data mining was invented, it is a process of extractingthe useful knowledge rules which are hidden from the large amount of data. And thedecision tree classification algorithm is the most widespread application of data mining,because of its simple basic principle, the faster calculation speed, and mining results whichare easily to understand, it is widely used in various fields.At the meantime, with the continuous expansion of college enrollment, the amount ofstudents increasing rapidly, so the amount of data associated with students also increasedsignificantly. But currently, the managements of most of the student evaluations are simpleentry calculation, some schools even still in the stage of paper managements. Thesemanagement approaches not only require a lot of manpower and material resources, butalso prone to error, missing data, etc. what worse, it only can achieve the most basicfunctions of inquiry statistics, but unable to discover the hidden data relationships and rules.However as we all known, the student evaluation as an important aspect of personneltraining, it is really necessary of using data mining to predict the results of the evaluationand find the association between evaluations to provide students training decisions.In this context, the paper focuses on improving C4.5, achieving student evaluationsystems and researching the application of the decision tree, main research contents can besummarized as follows:Above all, study and research C4.5decision tree classification algorithm, and improveit by analyzing the advantages and disadvantages for the algorithm and student evaluationsystem characteristics in the following two aspects: First, simplify formula, improve theefficiency of the algorithm, the second is adding balance factor to overcome some usersevaluation concerns the drawbacks distance from the root node in reality.Then, using AHP to build evaluation index and index weight, make sure to set areasonable and feasible evaluation index system. And according to these developing acollege student evaluation system, provide an effective tool which can give students someobjectives and scientific evaluations of students’comprehensive assessment calculations.Finally, introducing the applications of improved C4.5in the student evaluation system,and based on real students’ evaluation data for determining the object and the target of data mining, data collection and pre-treatment, constructing decision tree by improved C4.5andknowledge of the generate rules. Through the analysis of decision tree, you can understandthe relationship between important factories which impact the students’ assessment resultsand the various indicators, so can guide decision makers to adjust management measures atright time for achieving the purpose that improving students’ comprehensive assessmentcalculations.