Design And Implement Of Collection And Analysis System Of Students' Programming Behavior

In programming practice courses,students' understanding of programming language will be reflected in the complete behavior of programming,so programming behavior data has an important guiding role in improving students' programming ability.However,there is a general shortage of teaching assistants in programming courses,which leads to the lack of effective tracking of the student's programming process.At the same time,the mainstream tools on the market that support C++programming exercises are limited by the platform or experimental needs,and cannot meet the needs of collecting behavioral data as comprehensively as possible.To this end,we designed and implemented Students' Programming Behavior Collection and Analysis System,which can complete the collection and statistical analysis of students' programming behavior data and then dig deep into it.I am responsible for the research and development of the entire system.The main tasks include:1.Collecting student programming behaviors from multiple dimensions.By using VS extension technology to develop behavior collection plugin,our system is able to collect and record students' editing,compiling,debugging and testing behaviors in Visual Studio.2.Processing and statistically analysing the collected behavior logs.I analysed the collected behavior logs and stored the ayalysis result.Aiming for improving the quality of teaching,I conducted a fine-grained analysis of students'programming behavior from three perspectives:personal behavior,overall behavior,and abnormal behavior.3.Using behavioral data to explore the relationship between behavioral features and student performance.By using k-means to extract features of behavior data,I used decision tree,random forest,and SVM three classification prediction algorithms to classify and predict the programming ability of students,so as to explore the behavior features that are relatively related to programming ability,and achieve the purpose of providing students with effective suggestions for improving programming ability.At present,the system has been used for two semesters,and a total of about 16 million behavioral log records generated by 445 students in 13 effective exercises have been collected.The statistical analysis of these logs provides effective data support for teachers and teaching assistants in improving teaching content,and at the same time explores the differences in students' programming behavior.