.

although they were gathered for supervised learning purposes, you can try your algorithm on the CIFARs datasets

http://www.cs.toronto.edu/~kriz/cifar.html

other (rather) large datasets documented here (with performances related to supervised learning task)

http://rodrigob.github.io/are_we_there_yet/build/classification_datasets_results.html#53544c2d3130

.

That is great. Thank you. Do you know any disease dataset which has more than 1M samples or more?

.

medical data are way harder to find than images crawled from the web !

if you can make sense of their medical jargon (I can't ...), have a look at the simulator OSIM2 and some large simulated files available for download :

http://omop.fnih.org/OSIM2

(simulated, ok, but better than nothing ...)

.

not free (but not so expensive : the full KID database for about 100$ (student rate), for instance), pretty large and real data are available from HCUP

http://www.hcup-us.ahrq.gov/tech_assist/centdist.jsp

.

Dear Professor Fabrice,

Really thank you for your answers. I hope I will use your great knowledge in future. Also, I hope I will have a cooperation with you in future.

Yours sincerely,

Vahid Nouri

I want to use type-2 fuzzy clustering for a suitable dataset. Do you know any dataset which have uncertainity and this kind of clustering algorithm can deal with it?

you can find your clustering on UCI Machine Learning Repository