analyticPanelWise: Analytic panel-wise measures of predictive accuracy

In DudbridgeLab/multipred: Calculates measures of accuracy for risk predictors of multiple outcomes.

Analytic panel-wise measures of predictive accuracy

Description

Analytic calculation of panel-wise sensitivity, specificity, positive and negative predictive value, concordance and relative utility, under a multivariate liability threshold model.

Usage

analyticPanelWise(VL, VX, VLX = NULL, thresh = NULL, prev, nsample = NULL)

Arguments

VL	Variance-covariance matrix of liability. Must have 1 on diagonal.
VX	Variance-covariance matrix of predictors.
VLX	Cross-covariance matrix between liabilities and predictors. Entry on row i, column j, is covariance between liability i and predictor j. Diagonal entries are the liability variances explained for each trait.
thresh	Vector of risk thresholds for predicting an event. If NULL, which is the default, concordance is the only measure that can be calculated.
prev	Vector of prevalences, ie population risks, for each trait.
nsample	Number of random pairs of samples drawn when estimating concordance. If NULL, which is the default, concordance is not calculated. If 0, all possible pairs are drawn from the data.

Details

Panel-wise measures consider the prediction of at least one outcome to occur. At least one outcome that did occur must be predicted to occur. For example, panel-wise sensitivity is the probability that, for an individual in which at least one outcome did occur, the predicted risk exceeds the threshold for at least one of the outcomes that did occur. Panel-wise specificity is the probability that, for an individual in which at least one outcome did not occur, the predicted risk is lower than the threshold for all the outcomes that did not occur.

Panel-wise concordance is the probability that given one individual in which at least one outcome did occur, and another in which at least one did not occur, the maximum predicted risk over all outcomes that occurred in the former is higher than the maximum over all outcomes that did not occur in the latter. Note that under this definition an individual can be either concordant or discordant with itself. Concordance is calculated by randomly simulating nsample such pairs of individuals from the specified model.

Value

A list with the following components

sens Sensitivity

spec Specificity

PPV Positive predictive value

NPV Negative predictive value

C Concordance

RU Relative utility

Examples

# results will vary due to random sampling in computing multivariate integrals
attach(PRSdata)
analyticPanelWise(VL,VX,VX,thresh=prevalence,prev=prevalence,nsample=1e5)

# $sens
# [1] 0.6463497

# $spec
# [1] 0.0708455

# $PPV
# [1] 0.1081343

# $NPV
# [1] 0.9371735

# $C
# [1] 0.49142

# $RU
# [1] -0.31006

DudbridgeLab/multipred documentation built on Sept. 30, 2022, 2:24 a.m.