MPSEM (version 0.3-6)

lmutils: Linear modelling utility functions

Description

Utility functions to build linear models using Phylogenetic Eigenvector Maps as their features.

Usage

lmforwardsequentialsidak(y, x, object, alpha=0.05)
  lmforwardsequentialAICc(y, x, object)

Arguments

y

a response variable

x

descriptors to be used as auxiliary traits

object

a PEM class object

alpha

the threshold above which to stop adding variables

Value

Both functions return a lm class object.

Details

Function lmforwardsequentialsidak, performs a forward stepwise selection of the PEM eigenvectors until the familywise test of significance of the new variable to be included exceeds the threshold alpha. The familiwise type I error probability is obtained using the Holm-Sidak correction of the testwise probabilities, thereby correcting for type I error rate inflation due to multiple testing. lmforwardsequentialAICc carries out forward stepwise selection of the eigenvectors as long as the candidate model features a lower sample-size-corrected Akaike information criterion than the previous model. The final model should be regarded as overfit from the Neyman-Pearson (i.e. frequentist) point of view, but it is the model that minimizes information loss from the standpoint of information theory.

References

Burnham, K. P. & Anderson, D. R. 2002. Model selection and multimodel inference: a practical information-theoretic approach, 2nd ed. Springer-Verlag. xxvi + 488 pp.

Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scand. J. Statist. 6: 65-70.

Sidak, Z. 1967. Rectangular confidence regions for means of multivariate normal distributions. J. Am. Stat. Ass. 62, 626-633.

Examples

# NOT RUN {
## No example has yet been produced.
# }