Partial Area Estimation under Multi-Class Receiver Operating Characteristic Curve
Abstract
Receiver Operating Characteristic (ROC) curve is one of the most popular classification tools, which is widely used in a variety of fields, most notably in diagnostic medicine. AUC summarizes the entire curve, including the areas that may not be clinically relevant, especially those with high FPR. Generally, clinical studies do not encourage high FPR rates because living subjects are involved in it. In such cases, considering a portion of the ROC will be more meaningful, and such a portion of an ROC curve is named as pAUC. Literature has works on estimating the pAUC for bi-normal ROC curves in several scenarios. However, in most of the real-life scenarios, we cannot expect the normality always, so the existing model does not work in such cases. Also, when the data includes several sub-populations/multiclass (comprises more than two classes), the existing model will not work. To address these issues, in this paper, we have proposed a new methodology for estimating the pAUC of multiclass exponential ROC, which is used when we need to find the area under the curve that lies in a particular range of FPR and data, is of multi-class. The study is supported with simulation and real data studies.
References
- A. M. Lilienfeld; Some Limitations and Problems of Screening for Cancer, Cancer, 33 (S6), 1720–1724 (1974).
- D. K. McClish; Analyzing a Portion of the ROC curve. Medical Decision Making, 9(3), 190–
195 (1989).
- E. Hussain; The ROC Curve Model from Generalized-Exponential Distribution. Pakistan Journal of
Statistics and Operation Research, 7(2), 283–303 (2011).
- E. Hussain; The Bi-Gamma ROC Curve in a Straightforward Manner. Journal of Basic & Applied
Sciences, 8(2), 309–314 (2012).
- E. T. Lee and J. Wang; Statistical Methods for Survival Data Analysis. John Wiley & Sons, (2003).
- H. Ma; A.I. Bandos; H. E. Rockette and D. Gur; On Use of Partial Area under the ROC Curve for
Evaluation of Diagnostic Performance. Statistics in medicine, 32(20), 3449–3458 (2013).
- J. P. Egan; Signal Detection Theory and ROC Analysis. Academic Press, (1975).
- J. W. P. Tanner and J. A. Swets; A Decision-Making Theory of Visual Detection. Psychological Review, 61(6), 401-409(1954).
- L.B. Lusted; Signal Detectability and Medical Decision-making: Signal Detectability Studies Help
Radiologists Evaluate Equipment Systems and Performance of Assistants, Science, 171(3977), 1217–1219 (1971).
- L. E. Dodd and M. S. Pepe; Partial AUC Estimation and Regression. Biometrics, 59(3), 614–623 (2003).
- M. Goddard and I. Hinberg; Receiver Operator Characteristic (ROC) Curves and Non-Normal Data:
An Empirical Study. Statistics in Medicine, 9(3), 325–337 (1990).
- R. C. Elandt-Johnson; N. L. Johnson and M. Statistiker; Survival Models and Data Analysis. Wiley Online Library, (1980).
- R. G. Miller Jr; Survival Analysis. John Willey & Sons, (1981).
- R.V. Vardhan; S. Pundir and G. Sameera; Estimation of Area under the ROC Curve using Exponential and Weibull Distributions. Bonfring International Journal of Data Mining, 143-149 (2012).
- S. Balaswamy and R.V. Vishnu; An Anthology on Parametric ROC Models. Research & Reviews: Journal of Statistics, 532–546 (2016).
- S. Balaswamy and R.V. Vardhan; Estimation of the Area under the ROC Curve for Non-Normal Data. Communications in Statistics: Case Studies, Data Analysis and Applications, 8 (3), 393-406 (2022).
- W. J. Krzanowski and D. J. Hand, ROC Curves for Continuous Data, Crc Press, (2009).
- Y. He and M. Escobar, Nonparametric Statistical Inference Method for Partial Areas under Receiver Operating Characteristic Curves, with Application to Genomic Studies. Statistics in Medicine, 27 (25), 5291–5308 (2008).
- Y. Jiang; C. E. Metz and R. M. Nishikawa; A Receiver Operating Characteristic Partial Area Index
for Highly Sensitive Diagnostic Tests, Radiology, 201(3), 745–750 (1996).
Trion Studio
