Plagiarism Detection Tool for Programming Activity Logs
Name
Rene Kütt
Abstract
Plagiarism is a critical concern in academia, and educators need effective tools to detect and prevent plagiarism. Currently, most plagiarism detection tools use source code comparison, which is not potent against obfuscation methods used by students. This thesis presents a novel solution for detecting potential plagiarism in programming assignments using logs generated by Thonny IDE and an IntelliJ Platform plugin created as part of this thesis called PALG (Programming Activity Log Generator). A plagiarism detector has been incorporated into Thonny Log Analyser, a web application that processes logs created by Thonny IDE. PALA (Programming Activity Log Analyser) is a web application created by modifying Thonny Log Analyser to be compatible with log files produced by the IntelliJ Platform plugin. The plagiarism detection tool in the web application analyses the logs based on user-specified criteria such as run count, total time spent working, log file size, and pasted text to manually inserted text ratio.
The plagiarism detection tool also includes comparison functionalities that compare log files to each other or log files in different top folders, depending on the chosen analysis type. The comparison includes checking for duplicate files, identical texts pasted in different log files, source codes pasted in different log files and source code similarity detection. The similarity between log files is calculated using Dice's coefficient.
The purpose of the solution is to provide teachers with a quick and efficient overview of potential plagiarism in programming assignments. The solution is designed to work seamlessly with Moodle’s "Download all submissions" action, which produces a ZIP file that can be directly analysed by the web application. The solution's effectiveness is demonstrated through experimental evaluations, and the results indicate its potential to aid teachers in detecting plagiarism in programming assignments efficiently and effectively.
The plagiarism detection tool also includes comparison functionalities that compare log files to each other or log files in different top folders, depending on the chosen analysis type. The comparison includes checking for duplicate files, identical texts pasted in different log files, source codes pasted in different log files and source code similarity detection. The similarity between log files is calculated using Dice's coefficient.
The purpose of the solution is to provide teachers with a quick and efficient overview of potential plagiarism in programming assignments. The solution is designed to work seamlessly with Moodle’s "Download all submissions" action, which produces a ZIP file that can be directly analysed by the web application. The solution's effectiveness is demonstrated through experimental evaluations, and the results indicate its potential to aid teachers in detecting plagiarism in programming assignments efficiently and effectively.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Marina Lepp, Heidi Meier
Defence year
2023