fridge: Fridge

In khellton/fridge: Fridge: Focused ridge regression

Description

Function for calculating the fridge-ridge tuning parameter and prediction, with plot of MSE curve

Usage

fridge(X, y, x0, plug.in = c("OLS", "RLOOCV"), plot.curve = TRUE)

Arguments

X	Matrix with covariates of dimension n x p
y	Vector of outcomes of length n
x0	Vector of focus covariate of length p
plug.in	Character specifying the plug.in estimate, possibly 'OLS' (ordinary least squares) and 'RLO0CV' (ridge with leave-one-out cross-validation). The OLS option can only be used for p smaller than n.
plot.curve	Logical. Should the MSE curve and minimum value be ploted?

Details

The minimization routine is carried out with several different start values ranging from 1 to 10^7. If plot.curve = TRUE, the global minimum is ploted.

Value

Result, a list with the following

fridge.tuning	Numerical, the fridge tuning parameter
loocv.tuning	Numerical, the leave-one-out cross-validation tuning parameter
fridge.pred	Numerical, the fridge prediction
loocv.pred	Numerical, the leave-one-out cross-validation prediction

Examples

fridge()

khellton/fridge documentation built on May 20, 2019, 2:26 p.m.