Omar, Luma and Ivrissimtzis, Ioannis (2019) 'Using theoretical ROC curves for analysing machine learning binary classifiers.', Pattern recognition letters., 128 . pp. 447-451.
Most binary classifiers work by processing the input to produce a scalar response and comparing it to a threshold value. The various measures of classifier performance assume, explicitly or implicitly, probability distributions Ps and Pn of the response belonging to either class, probability distributions for the cost of each type of misclassification, and compute a performance score from the expected cost. In machine learning, classifier responses are obtained experimentally and performance scores are computed directly from them, without any assumptions on Ps and Pn. Here, we argue that the omitted step of estimating theoretical distributions for Ps and Pn can be useful. In a biometric security example, we fit beta distributions to the responses of two classifiers, one based on logistic regression and one on ANNs, and use them to establish a categorisation into a small number of classes with different extremal behaviours at the ends of the ROC curves.
|Full text:||Publisher-imposed embargo until 04 October 2020. |
(AM) Accepted Manuscript
Available under License - Creative Commons Attribution Non-commercial No Derivatives.
File format - PDF (571Kb)
|Publisher Web site:||https://doi.org/10.1016/j.patrec.2019.10.004|
|Publisher statement:||© 2019 This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/|
|Date accepted:||03 October 2019|
|Date deposited:||22 October 2019|
|Date of first online publication:||04 October 2019|
|Date first made open access:||04 October 2020|
|Social bookmarking:||Export: EndNote, Zotero | BibTex|
|Look up in GoogleScholar | Find in a UK Library|