Compositional (version 1.0)

glmpcr.tune: Tuning the principal components in the GLM

Description

Tuning the number of principal components in the generalised linear models.

Usage

glmpcr.tune(y, x, M = 10, maxk = 10, oiko = "binomial", seed = FALSE, ncores = 2)

Arguments

y

A real valued vector.

x

The predictor variables, they have to be continuous.

M

The number of folds in the cross validation.

maxk

The maximum number of principal components to check.

oiko

The type of regression to be used, "binomial" if you have binary data or "poisson" if you have count data.

seed

If seed is TRUE the folds will always be the same.

ncores

The number of cores to use. If more than 1, parallel computing takes place.

Value

• A list including:
• mspe
  A vector with the mean squared error of prediction (MSPE), each value corresponds to a number of principal components.
• k
  The number of principal components which minimizes the MSPE.
• performance
  The bias corrected lowest value of the MSPE along with the estimated bias via the Tibshirani and Tibshirani (2009) suggestion.

Details

Cross validation is performed to select the optimal number of principal components in the GLM. This is used by alfapcr.tune.

References

Jolliffe I.T. (2002). Principal Component Analysis. Tibshirani and Tibshirani (2009). A bias correction for the minimum error rate in cross-validation. The Annals of Applied Statistics, 3(1):822-829.

See Also

pcr.tune, glm.pcr, alfa.pcr, alfapcr.tune

Examples

library(MASS)
x <- fgl[, 2:9]
y <- rpois(214, 10)
glmpcr.tune(y, x, M=10, maxk = 20, oiko = "poisson", seed = FALSE, ncores = 1)