predict.MTPS: Make predictions from a "MTPS" model

In ubcxzhang/MTPS: Multi-Task Prediction using Stacking Algorithms

Description

This function makes predictions from a revised stacking model.

Usage

## S3 method for class 'MTPS'
predict(object, newdata, ...)

Arguments

object	A fitted object from "MTPS"
newdata	Matrix of new predictors at which predictions are to be made
...	additional arguments affecting the predictions produced

Value

The predicted value from new predictors.

Examples

data("HIV")
set.seed(1)
xmat <- as.matrix(XX)
ymat <- as.matrix(YY)
id <- createFolds(rowMeans(XX), k=5, list=FALSE)
training.id <- id != 1
y.train <- ymat[training.id, ]
y.test  <- ymat[!training.id, ]
x.train <- xmat[training.id, ]
x.test  <- xmat[!training.id, ]
# Cross-Validation Residual Stacking
fit.rs <- MTPS(xmat = x.train, ymat = y.train,
  family = "gaussian",cv = FALSE, residual = TRUE,
  method.step1 = rpart1, method.step2 = lm1)
pred.rs <- predict(fit.rs, x.test)

ubcxzhang/MTPS documentation built on Jan. 29, 2020, 9:12 p.m.