Research And Application Of Student Performance Prediction Algorithm Based On C5.0 Decision Tree Algorithm

In early 2020,with the outbreak and spread of the new crown epidemic in my country,the teaching of universities across the country was severely affected.At this time,replacing the traditional teaching mode with online teaching mode has become an inevitable choice under the epidemic.At present,there are many network teaching systems that provide teachers to teach courses on the Internet and let students learn on the Internet.But network teaching is not as intuitive as classroom teaching,which leads to the teachers' judgment of students' learning situation is not as accurate as classroom teaching.So how to effectively analyze the effect of online teaching has always been a hot issue in teaching research.In the case of network teaching,in order to make teachers observe all the students' learning behaviors and learning conditions in real time,this study proposes an integrated analysis algorithm.The students' learning behavior can be analyzed by using the existing learning record files through multiple observation dimensions,and the students who may not reach the expected teaching objectives can be predicted by C5.0 decision tree algorithm,so that teachers can Timely give appropriate teaching assistance to achieve teaching objectives.The research content of this paper mainly includes the following two aspects:1.Based on C5.0 decision tree,an algorithm for predicting students' learning achievement is proposed.The algorithm determines the input and output variables of decision tree by the data table of students' learning behavior factors.Then,all samples are randomly divided into three data sets,which are used for training,testing and verification.According to this analysis,the characteristics of students' learning behavior can provide the algorithm basis for student performance prediction.C5.0 decision tree algorithm has established three databases related to learning record analysis: student learning record database,student learning type database,and teaching decision support database.And after the analysis of the results of the decision-making rules,it is divided into three parts: to find out the decision rules between the attributes of learning behavior and learning effectiveness of students in the fifth and sixth semester courses,and to explain the decision rules.In order to verify the decision rules,the same course of different semester is used as the object of verification,and the data of the fifth semester,sixth semester and seventh semester are used as training data and test data respectively.The training error rate and test error rate in decision-making analysis are used as the verification indicators to try to find the best prediction by different time units at the time point,different grades are used to find out the best classification method of grades.The analysis objects are divided into the whole students,high-grade students and low-level students.2.Based on C5.0 decision tree,the student learning achievement prediction decision tree is analyzed and discussed,and the specific educational guidance is given for the data analysis results.This study deals with these heterogeneous and large number of learning records,from the relationship between curriculum attributes and decision rules under different courses,the relationship between time points and decision rules in different time units,the relationship between classification methods of different grades and decision rules,and the relationship between time points and prediction effect of high score and low score in different time units,classification of different grades and high level The relationship between score and low score prediction results can analyze the learning record information,classify the students' learning behavior,and finally achieve the purpose of actively predicting the possible learning effect according to the current learning situation of students.