Err.sgl: Error Rates

In sglOptim: Generic Sparse Group Lasso Solver

Description

Compute and return the root-mean-square error for each model. This method is only intended for testing.

The root-mean-square error (RMSE) is

\frac{1}{K}∑_{i = 1}^K √{\frac{1}{N}∑_{j=1}^N Y_{ji}-(X\hat β)_{ji}}

RMSE is the default error.

Usage

## S3 method for class 'sgl'
Err(object, data = NULL, response = object$Y.true, ...)

Arguments

object	a lsgl object.
data	a design matrix (the X matrix).
response	a matrix of the true responses (the Y matrix).
...	ignored.

Value

a vector of errors.

Author(s)

Martin Vincent

sglOptim documentation built on May 8, 2019, 1:02 a.m.