Implements multiple performance measures for supervised learning. Includes over 40 measures for regression and classification. Additionally, meta information about the performance measures can be queried, e.g. what the best and worst possible performances scores are. Internally, checkmate is used to check arguments efficiently - no other runtime dependencies.
The function reference gives an encompassing overview over implemented measures.