Abstract
The Receiver Operating Characteristic (ROC) curve analysis and the resulting plot can be used as a tool to select optimal models of possibility and to discard those of inferior quality from the cost of context (or class distribution). Presently, this type of analysis is used in a variety of fields from the medical community, bioinformatics, military and finance. There is a variety of software packages available for ROC analysis, and this analysis will focus on those specific of R and open source. The chosen packages were: ROCR, Verification, caTools, Comp2ROC, and Epi available on CRAN, and the ROC library from Bioconductor. This work intends to make a comparative analysis of the main characteristics of these R packages.
Author supplied keywords
Cite
CITATION STYLE
Da Cunha, D. F., & Braga, A. C. (2017). Receiver operating characteristic (ROC) packages comparison in R. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10405, pp. 545–559). Springer Verlag. https://doi.org/10.1007/978-3-319-62395-5_37
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
