Editing Receiver operating characteristic (section)

==Criticisms==
[[File:ROC curve example highlighting sub-area with low sensitivity and low specificity.png|thumb|Example of receiver operating characteristic (ROC) curve highlighting the area under the curve (AUC) sub-area with low sensitivity and low specificity in red and the sub-area with high or sufficient sensitivity and specificity in green.<ref name="Chicco Jurman 2023 p. "/>]]

Several studies criticize certain applications of the ROC curve and its area under the curve as measurements for assessing binary classifications when they do not capture the information relevant to the application.<ref name="Muschelli 2019 pp. 696–708">{{cite journal | last=Muschelli | first=John | title=ROC and AUC with a binary predictor: a potentially misleading metric | journal=Journal of Classification | publisher=Springer Science and Business Media LLC | volume=37 | issue=3 | date=2019-12-23 | issn=0176-4268 | doi=10.1007/s00357-019-09345-1 | pages=696–708| pmid=33250548 | pmc=7695228 }}</ref><ref name="Chicco Jurman 2023 p. ">{{cite journal | last1=Chicco | first1=Davide | last2=Jurman | first2=Giuseppe | title=The Matthews correlation coefficient (MCC) should replace the ROC AUC as the standard metric for assessing binary classification | journal=BioData Mining | publisher=Springer Science and Business Media LLC | volume=16 | issue=1 | date=2023-02-17 | issn=1756-0381 | doi=10.1186/s13040-023-00322-4 | page=4| doi-access=free | pmid=36800973 | pmc=9938573 | hdl=10281/430042 | hdl-access=free }}</ref><ref name="Lobo Jiménez-Valverde Real 2008 pp. 145–151">{{cite journal | last1=Lobo | first1=Jorge M. | last2=Jiménez-Valverde | first2=Alberto | last3=Real | first3=Raimundo | title=AUC: a misleading measure of the performance of predictive distribution models | journal=Global Ecology and Biogeography | publisher=Wiley | volume=17 | issue=2 | year=2008 | issn=1466-822X | doi=10.1111/j.1466-8238.2007.00358.x | pages=145–151| bibcode=2008GloEB..17..145L }}</ref><ref name="Halligan Altman Mallett 2015 pp. 932–939">{{cite journal | last1=Halligan | first1=Steve | last2=Altman | first2=Douglas G. | last3=Mallett | first3=Susan | title=Disadvantages of using the area under the receiver operating characteristic curve to assess imaging tests: A discussion and proposal for an alternative approach | journal=European Radiology | publisher=Springer Science and Business Media LLC | volume=25 | issue=4 | date=2015-01-20 | issn=0938-7994 | doi=10.1007/s00330-014-3487-0 | pages=932–939| doi-access=free | pmid=25599932 | pmc=4356897 }}</ref><ref name="Berrar Flach 2011 pp. 83–97">{{cite journal | last1=Berrar | first1=D. | last2=Flach | first2=P. | title=Caveats and pitfalls of ROC analysis in clinical microarray research (and how to avoid them) | journal=Briefings in Bioinformatics | publisher=Oxford University Press (OUP) | volume=13 | issue=1 | date=2011-03-21 | issn=1467-5463 | doi=10.1093/bib/bbr008 | pages=83–97| doi-access=free | pmid=21422066 }}</ref>

The main criticism to the ROC curve described in these studies regards the incorporation of areas with low sensitivity and low [[specificity (tests)|specificity]] (both lower than 0.5) for the calculation of the total area under the curve (AUC).,<ref name="Lobo Jiménez-Valverde Real 2008 pp. 145–151" /> as described in the plot on the right.

According to the authors of these studies, that portion of area under the curve (with low sensitivity and low specificity) regards confusion matrices where binary predictions obtain bad results, and therefore should not be included for the assessment of the overall performance.
Moreover, that portion of AUC indicates a space with high or low confusion matrix threshold which is rarely of interest for scientists performing a binary classification in any field.<ref name="Lobo Jiménez-Valverde Real 2008 pp. 145–151" />

Another criticism to the ROC and its area under the curve is that they say nothing about precision and negative predictive value.<ref name="Chicco Jurman 2023 p. "/>

A high ROC AUC, such as 0.9 for example, might correspond to low values of precision and negative predictive value, such as 0.2 and 0.1 in the [0, 1] range.
If one performed a binary classification, obtained an ROC AUC of 0.9 and decided to focus only on this metric, they might overoptimistically believe their binary test was excellent. However, if this person took a look at the values of precision and negative predictive value, they might discover their values are low.

The ROC AUC summarizes sensitivity and specificity, but does not inform regarding precision and negative predictive value.<ref name="Chicco Jurman 2023 p. "/>