modelPredict_VS: Risk prediction on the test set, variable selection version

Description Usage Arguments Value Examples

Description

For a fixed threshold value τ, modelPredict_VS predicts the outcome Y for subjects in the test set. This function also outputs the cost associated with the prediction procedure. This function is used when the baseline covariates Z are high-dimensional.

Usage

1
modelPredict_VS(list_fpcaFit, list_cvfit, Xtest, Ztest, startT, tau)

Arguments

list_fpcaFit

Obtained FPCA decomposition from modelFit_VS.

list_cvfit

Obtained elastic net logistic regression from modelFit_VS.

Xtest

Longitudinal biomarker data for subjects in the test set, matrix of testn by nTotal. Missing values are denoted by NA.

Ztest

Other baseline covariates for subjects in the test set.

startT

Time of the first prediction, denoted by t_1 in the manuscript. For instance, if the time grids are {0,1/60,2/60,...,1}, then startT = 25 means that the first prediction is made at t = 24/60.

tau

The threshold value τ.

Value

final.label

Predicted outcome Y for subjects in the test set, vector of length testn.

avg.cost

Average cost when we applied this prediction procedure to the test set.

cost

Cost for each subject, vector of length testn. For some subjects, we make a definite decision early. For others, we follow up with a long period of time. Hence the cost is different for each individual.

Examples

1
# see the example from function reinforced_VS.

reinforcedPred documentation built on May 2, 2019, 4:17 a.m.