Hi Stevan Cakic You can jump on this research articles if you can scrape some feeds from them.

Article Identifying neuropsychiatric disorders using unsupervised cl...

Article Multi-View Image Clustering Based on Sparse Coding and Manif...

I perceive the first link will open certain doors in your search.

All the best

Interesting question. Following the discussion.

Akpofure A Enughwure thank you. I will check these articles.

Googling "Clustering python" will get you plenty of information on modern methods of clustering. Here's my first hit:

https://machinelearningmastery.com/clustering-algorithms-with-python/

Dear Alan Parker, thanks. I googled of course and saw some of the very useful articles, but this is a concrete use-case. Also, I posted this article in the hope to find someone with close interests.

The main goal of this research is to group similar programming assignments in clusters, so that can help me as a teaching assistant to know if some of the students cheating on tests or homework.

I wonder if clustering is the best approach. I hope that cheating is rare, so wouldn't outlier detection be a better way?

Alan Parker nice idea, maybe two approaches should be discussed here:

- when cheating is rare (for example exams and tests);

- when cheating is common (sad but true, this is more common in programming homeworks then the first approach)

Preprint Probability and metric on finite sets. Partitions, orderline...

Thank you. I'm thinking about converting code to AST and then try some of the clustering methods. It seems to me that can be the right path.