ThresholdROC: Optimum Threshold Estimation Tools for Continuous Diagnostic Tests in R
Autor: | Natàlia Pallarès, Sara Perez-Jaume, Josep L. Carrasco, Konstantina Skaltsa |
---|---|
Přispěvatelé: | Universitat de Barcelona |
Jazyk: | angličtina |
Rok vydání: | 2017 |
Předmět: |
0301 basic medicine
Statistics and Probability Computer science 030230 surgery computer.software_genre 03 medical and health sciences diagnostic tests 0302 clinical medicine Diagnòstic Diagnosis Parameter estimation Threshold estimation ROC curve threshold estimation cost function R package bootstrap lcsh:Statistics lcsh:HA1-4737 Parametric statistics Estimation Sampling (statistics) Diagnostic test R (Llenguatge de programació) Function (mathematics) Confidence interval Estimació d'un paràmetre 030104 developmental biology Sample size determination R858-859.7 Computer applications to medicine. Medical informatics R (Computer program language) Data mining Statistics Probability and Uncertainty computer Software |
Zdroj: | Journal of Statistical Software, Vol 82, Iss 1, Pp 1-21 (2017) Recercat. Dipósit de la Recerca de Catalunya instname Journal of Statistical Software; Vol 82 (2017); 1-21 Dipòsit Digital de la UB Universidad de Barcelona |
ISSN: | 1548-7660 |
Popis: | We introduce an R package that estimates decision thresholds in diagnostic settings with a continuous marker and two or three underlying states. The package implements parametric and non-parametric estimation methods based on minimizing an overall cost function, as well as confidence interval estimation approaches to account for the sampling variability of the cut-off. Further features of the package include sample size determination and estimation of diagnostic accuracy measures. We used randomly generated data and two real datasets to illustrate the capabilities and characteristics of the package. |
Databáze: | OpenAIRE |
Externí odkaz: |