multisite.accuracy: Estimation of Accuracy in Multisite Machine-Learning Models

The effects of the site may severely bias the accuracy of a multisite machine-learning model, even if the analysts removed them when fitting the model in the 'training set' and when applying the model in the 'test set'. This simple R package estimates the accuracy of a multisite machine-learning model unbiasedly as described in (Solanes et al, Psychiatry Research: Neuroimaging 2021, in Press). It currently supports the estimation of sensitivity, specificity, balanced accuracy, the area under the curve, correlation, mean squarer error, and hazard ratio for binomial, gaussian, and survival (time-to-event) outcomes.

Version: 1.0
Imports: AROC, coxme, lme4, lmerTest, logistf, metafor, pROC, survival
Published: 2021-05-28
Author: Joaquim Radua
Maintainer: Joaquim Radua <quimradua at gmail.com>
License: GPL-3
NeedsCompilation: no
CRAN checks: multisite.accuracy results

Downloads:

Reference manual: multisite.accuracy.pdf
Package source: multisite.accuracy_1.0.tar.gz
Windows binaries: r-devel: multisite.accuracy_1.0.zip, r-release: multisite.accuracy_1.0.zip, r-oldrel: multisite.accuracy_1.0.zip
macOS binaries: r-release (arm64): multisite.accuracy_1.0.tgz, r-release (x86_64): multisite.accuracy_1.0.tgz, r-oldrel: multisite.accuracy_1.0.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=multisite.accuracy to link to this page.