A machine learning algorithm to stop cheating in the classroom.
Unfortunately plagiarism is a very serious problem in computer science courses. Mimir Classroom's configurable plagiarism detection system acts as a strong deterrent for those tempted to cheat and catches those that do anyway.
Three Levels of Detection
Current Class Detection
Analyzes each student’s submission against the rest of the class to identify cases of students sharing work.
Historical Class Detection
When a programming project has been used before we will automatically check each student’s submission against all previous submissions, including previous semesters.
Web Detection
Checks student submissions for similarities from websites like Stack Overflow and RosettaCode.
More than a Simple Diff
Using complex structural, user patterns, and post compilation analysis, Mimir can catch many of the tricks employed by students to get away with cheating.
Machine Learning
The Mimir Plagiarism Algorithm is constantly learning. It learns from feedback and quickly adapts to new cheating methods and different project types.