multipred: multipred

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

multipred

Description

Package for calculating measures of accuracy for risk predictors of multiple outcomes.

Details

Accuracy can be evaluated in four senses: outcome-wise, joint, and panel-wise (weak sense and strong sense). For convenience the weak panel-wise sense is also called "screening", and the strong panel-wise sense simply "panel-wise". In each sense, accuracy can be measured empirically, within data sets given as input, or theoretically, given parameters of an underlying multivariate liability threshold model.

Throughout the documentation, an "outcome" means one of several binary variables observed in an individual, and an outcome "occurs" when the variable has the positive state.

Outcome-wise measures calculate standard univariate measures of accuracy over all outcomes and individuals.

Joint measures consider the prediction of all outcomes occuring simultaneously within an individual.

Screening measures consider the prediction of at least one outcome occuring within an individual. It is not necessary that the predicted outcomes are the same ones that actually occur.

Panel-wise measures consider the prediction of at least one outcome occuring within an individual. There must be at least one predicted outcome that actually occurs.

Functions

outcomeWise

joint

screening

panelWise

analyticOutcomeWise

analyticJoint

analyticScreening

analyticPanelWise

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