| dc.contributor.author |
Santos Costa, Vitor |
|
| dc.contributor.author |
Page, David |
|
| dc.contributor.author |
Davis, Jesse |
|
| dc.contributor.author |
Boyd, Kendrick |
|
| dc.date.accessioned |
2012-07-12T17:35:08Z |
|
| dc.date.available |
2012-07-12T17:35:08Z |
|
| dc.date.issued |
2012-05-30 |
|
| dc.identifier.citation |
TR1772 |
en |
| dc.identifier.uri |
http://digital.library.wisc.edu/1793/61736 |
|
| dc.description.abstract |
Precision-recall (PR) curves and the areas under them are widely used to summarize machine learning results, especially for data sets exhibiting class skew. They are often used analogously to ROC curves and the area under ROC curves. It is known that PR curves vary as class skew changes. What was not recognized before this paper is that there is a region of PR space that is completely unachievable, and the size of this region depends only on the skew. This paper precisely characterizes the size of that region and discusses its implications for empirical evaluation methodology in machine learning. |
en |
| dc.description.provenance |
Submitted by Jody Hoesly (jhoesly@wisc.edu) on 2012-07-12T17:35:08Z
No. of bitstreams: 1
TR1772 prspace_techreport.pdf: 418934 bytes, checksum: 41090e9a1c6fc0d39d54a4c996a11839 (MD5) |
en |
| dc.description.provenance |
Made available in DSpace on 2012-07-12T17:35:08Z (GMT). No. of bitstreams: 1
TR1772 prspace_techreport.pdf: 418934 bytes, checksum: 41090e9a1c6fc0d39d54a4c996a11839 (MD5)
Previous issue date: 2012-05-30 |
en |
| dc.publisher |
University of Wisconsin-Madison Department of Computer Sciences |
en |
| dc.subject |
F1 score |
en |
| dc.subject |
precision-recall curves |
en |
| dc.title |
Unachievable Region in Precision-Recall Space and Its Effect on Empirical Evaluation |
en |
| dc.type |
Technical Report |
en |
| dc.contributor.affiliation |
University of Wisconsin-Madison Department of Computer Sciences |
|
