Regression/Classification info

[ogencoglu]

Hi,

It would be nice to have a 3rd column for data() output indicating whether the dataset can be used for regression or classification problems.

[iamaziz]

Hi @ogencoglu sounds like a cool idea, thanks. Any thought on how to approach clustering them ?

[ogencoglu]

I think it is just manual work. Not all datasets may be suitable for this but many machine learning people search for datasets to try their algorithms/implementations in a smaller scale before going to well-known benchmark datasets.

[mynameisvinn]

agreed it'd be nice to filter for regression or classification, but dont see how you could properly categorize datasets.

a regression dataset could be a classification dataset, and vice versa, depending on your preprocessing strategy (eg binning) and target feature.

for example, the canonical iris dataset, used for classification, could be viewed as regression too.

[ogencoglu]

My idea was something similar to UCI data repo: http://archive.ics.uci.edu/ml/datasets.html

The column can be "Default Task". Some datasets may have even both Classification and Regression.