6
$\begingroup$

I'm implementing a decision tree algorithm, and I'd like to get a feel for how it performs relative to other implementations. Can anyone recommend popular datasets for training and testing decision tree algorithms? I've found some resources like this, but I'm not sure widely used it is.

$\endgroup$
3
$\begingroup$

I would recommend the UCI repository you're mentioning. It has been around for quite some time, contains many data sets and is frequently referenced in scientific publications.

$\endgroup$
3
$\begingroup$

You're on the right track with the UCI repository. MLcomp is another great resource, and will automatically score your algorithm on multiple datasets.

$\endgroup$
2
$\begingroup$

You could try having a look at the datasets from Kaggle competitions at kaggle.com. Some require a fair degree of pre-processing, but there are some relatively 'clean' datasets there. You can see how your algorithm performs by submitting predictions to either current or past competitions and see how well it performs relatively to other participants.

$\endgroup$
1
$\begingroup$

Thought I should mention milksets, a Python wrapper around some UCI datasets. It appears to have 7 of the datasets, and produces them as a simple 2D Numpy array.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.