predict.MLGL: Predict fitted values from a 'MLGL' object

In MLGL: Multi-Layer Group-Lasso

Predict fitted values from a MLGL object

Description

Predict fitted values from a MLGL object

Usage

## S3 method for class 'MLGL'
predict(object, newx = NULL, s = NULL, type = c("fit", "coefficients"), ...)

Arguments

object	MLGL object
newx	matrix with new individuals for prediction. If type="coefficients", the parameter has to be NULL
s	values of lambda. If NULL, use values from object
type	if "fit", return the fitted values for each values of s, if "coefficients", return the estimated coefficients for each s
...	Not used. Other arguments to predict.

Value

A matrix with fitted values or estimated coefficients for given values of s.

Author(s)

original code from gglasso package Author: Yi Yang <[e-mail]>, Hui Zou <[e-mail]>

function inspired from predict function from gglasso package by Yi Yang and Hui Zou.

See Also

MLGL

Examples

X <- simuBlockGaussian(n = 50, nBlock = 12, sizeBlock = 5, rho = 0.7)
y <- drop(X[, c(2, 7, 12)] %*% c(2, 2, -1)) + rnorm(50, 0, 0.5)

m1 <- MLGL(X, y, loss = "ls")
predict(m1, newx = X)
predict(m1, s=3, newx = X)
predict(m1, s=1:3, newx = X)

MLGL documentation built on March 31, 2023, 9:32 p.m.