yhat_step: Backward Stagewise Regression with AIC or BIC
In gencve: General Cross Validation Engine

Description Usage Arguments Value Author(s) Examples

Fits a subset regression model using backward stagewise regression to training data and computes the predictions for the test data.

1	yhat_step(dfTrain, dfTest, ic = c("BIC", "AIC"))

`dfTrain`	Data frame for training data. Last column must be the output variable.
`dfTest`	Data frame for test data. Last column must be the output variable.
`ic`	Information criterion to use to select the number of components. Default is BIC.

The predictions for the test sample

A. I. McLeod

Xy <- prostate
X <- prostate[,-9]
y <- prostate[,9]
n <- length(y)
d <- 10
set.seed(777513)
iTe <- sample(n, size=d)
iTr <- (1:n)[!match(1:n, iTe, nomatch = 0) > 0]
trdf <- data.frame(X[iTr,], y=y[iTr]) #X, y already defined
tedf <- data.frame(X[iTe,], y=y[iTe])
yhat_step(trdf, tedf)