Donation Policy:

The construction of this repository is an ongoing process. The majority of entries in the repository were contributed by machine learning researchers outside of UCI. If you have or know of an entry that we should include here, please let us know. We are always interested in donations.

The best way to donate data sets is to fill out our web form which will allow you to upload your data file into our repository. The metadata that we request is useful for efficient faceted search of data sets:

Donate a Data Set!

Alternatively, donations may be made by doing an anonymous ftp of donated files to the /incoming directory at ftp.ics.uci.edu. After ftping your files, please send an email to ml-repository@ics.uci.edu. You may also send us the data as an email attachment if the data set is small, or you may point us to a place on the web that contains the data that you are donating.

While we accept entries with minimal documentation, we prefer that the donor do their best at documenting the files in order to provide maximum benefit to the machine learning community. The file Doc-Requirements describes the standard documentation format.

We also prefer that the data have a standard format. For databases of instances, the standard format is one instance per line, no spaces, comma separated attribute values, and missing values denoted by "?". For other types of data, please use your best judgement.

Thank you for your donations! You are helping to make machine learning a strong and vital research area.

Supported By:

 In Collaboration With: