predict.sJIVE: Prediction for sJIVE

In enorthrop/sup.r.jive: Supervised JIVE Methods: JIVE Predict, sJIVE, and sesJIVE

Prediction for sJIVE

Description

Predicted values based on the an sJIVE model.

Usage

## S3 method for class 'sJIVE'
predict(object, newdata, threshold = 0.001, max.iter = 2000, ...)

Arguments

object	An object of class "sJIVE", usually a fitted sJIVE model.
newdata	A list of matrices representing the new X datasets.
threshold	The threshold used to determine convergence of the algorithm.
max.iter	The maximum number of iterations for each instance of the sJIVE algorithm.
...	further arguments passed to or from other methods.

Details

predict.sJIVE calculates predicted values for newdata based on the fitted model. The function first calculates the joint and individual score matrices for newdata. Note that the fitted model's loadings and coefficients are treated as known and will not get re-calculated. Once the new score matrices are obtained, the linear prediction model will be evaluated using the new scores as the data matrix.

Value

A list of the following components is returned:

Ypred	The fitted Y values.
S_J	A matrix capturing the joint scores of newdata.
S_I	A list containing matrices that capture the individual scores of newdata.
iterations	The number of iterations needed to reach convergence.
error	The error value at which the model converged.

Examples

train.x <- list(matrix(rnorm(300), ncol=20),matrix(rnorm(200), ncol=20))
train.y <- rnorm(20)
test.x <- list(matrix(rnorm(600), ncol=40),matrix(rnorm(400), ncol=40))
train.fit <- sJIVE(X=train.x,Y=train.y,rankJ=1,rankA=c(1,1),eta=0.5)
test.fit <- predict(train.fit, newdata = test.x)

enorthrop/sup.r.jive documentation built on Nov. 18, 2022, 6:01 p.m.