A plagiarized code may be either an exact copy of the original code, or a variant, applying various textual transformations such as;
- Changing the order of operands/operators in expressions,
- Changing data types,
- Adding redundant statements or variables, and
- Changing comments,
- Changing white space ,
- Renaming and reordering identifiers,
- Reordering and replacing code blocks,
- Reordering statements within code blocks,
I would like to make a research on the subject of software code plagiarism checking.
Is there any publicly available benchmark dataset?
Hi Deniz,
try this software:
http://jplag.ipd.kit.edu/
http://theory.stanford.edu/~aiken/moss/
Hi Janusz,
Thank you but, I think, there is a misunderstanding.
I'm looking for the dataset not software.
Hi Deniz,
how large dataset you need? I have homeworks from my students.
Hi Janusz,
Since we will first do preprocess files, any number is welcome :)
Language is also not important. There is one thing to pay attention is the complexity level of homework. For example, homeworks including functions/procedures, data structures (arrays, structs, classes) are appropriate.
Best regards,
Deniz.
Would you mind discussing the kind of algorithm you are using for Code Plagiarism Detection? Is this using the longest sub-sequence or some other algorithm?
You can check this out:
http://users.dsic.upv.es/grupos/nle/soco/
SoCo or Detection of SOurce COde Re-use is being offered as a track in Forum for Information Retrieval Evaluation (FIRE). The track will again be offered this year in FIRE. You can contact the organizers if they can make last years data available, or else you can participate in this year's SoCo in FIRE and get the dataset.
- Badges
- Science topic