Similarity Assessment

This app still has some problems under Linux, possibly in the context of the R-Python interface.

A Shiny app for automatic grading of assignments.

First, upload some files in one batch to the shiny app by clicking browse. Please do not upload more than three files at once.
You can download an archive of sample files.

Next, pick the type of similarity. Symmetric similarity measures how much of the code between the sample solution (or other student file) and student file is the same. Amount of solution in submission measures the amount of code similar to the sample solution found a student file. Finally, amount of submission in solution measures the amount of code similar to a student file found in the sample solution. All of these comparisons are pairwise.

After the choice in similarity, you can choose to compare all files to each other, or to compare to a sample solution. If you compare all files to each other, the app can be used to test for plagiarism. The plagiarism decision is shown in the decision column, 0 for not plagiarized, 1 for possible plagiarism. The system decides if a code is plagiarized if a code is more than 70% similar to another work.

If compared to a sample solution, the amount of similarity may be used as a grading tool. The underlying assumption is that code that is more similar to a sample solution is of higher quality, and therefore should have a higher grade.

QHELP is co-funded by the Erasmus+ Programme of the European Commission.

This server is powered by RStudio's Shiny Server Open Source.

© 2022 Sohum Bhatt, Wim Van den Noortgate & Francis Tuerlinckx, KU Leuven, Belgium

Legal Disclaimer | Privacy Information