Dear Dr. Zenon Gniazdowski

As you know, PCA is a single classifier. Therefore, it can utilize each class of any data set as target data. But its performance is directly related to data distribution. To introduce several data sets, the addresses of  some data repositories are added to this letter. I hope these may help you.

Best regards,

Reza Sadeghi

http://keel.es/datasets.php

http://archive.ics.uci.edu/ml/datasets.html

http://db.csail.mit.edu/labdata/labdata.html

Thank you very much.

I know these repositories. I have used the resources of the UCI repository.

My problem is not finding a repository, but finding a set of correlated data. I would like to find sets of numbers, containing the correlated data, similar to the Iris set.

This tutorial may help,

http://www.kdnuggets.com/2016/06/nutrition-principal-component-analysis-tutorial.html

The data from United States Department of Agriculture has been used to find the best way to differentiate food items.

since you are applying PCA..you can take datasets with many feature sets....i applied pca on this bank marketing dataset

https://archive.ics.uci.edu/ml/datasets/bank+marketing

I do not exclude that one day I will return to the problem and I will also analyze your data. Thank you for your suggestion!

However, I regret that I did not have time to use your proposal before the end of the article: https://www.researchgate.net/publication/328345955_On_the_clustering_of_correlated_random_variables

I do not rule out that I will return to the problem and I will also analyze your data.