rmdr.fast {Rmdr}R Documentation

R-Multifactor Dimensionality Reduction Function

Description

rmdr.fast(R function based on C programs of the MDR approach ): Detection and Characterization of Gene-Gene Interactions and Gene-Environment Interactions in Genetic and Epidemiological Studies. MDR is a nonparametric and genetic model-free alternative to logistic regression for detecting and characterizing nonlinear interactions among discrete genetic and environmental attributes.

Usage

rmdr.fast(data, cvk, nbr, ratio = NULL, randomize = FALSE)

Arguments

data the data to be used. The fisrt column describes the class variable (0 for unaffected and 1 for affected). The other columns describe the attributes with values 0,1,2.
cvk The number of cross-validation > 1
nbr the number of attribute combinations to be considered
ratio the threshold that is used to pool attribute levels into two groups
randomize randomly order the data

Value

A matrix giving the best combination of loci, the misclassification error and the prediction error.

Author(s)

Mounir Aout, contributions from C. Wachter

References

Hahn LW, Ritchie MD, Moore JH.: Multifactor dimensionality reduction software for detecting gene-gene and gene-environment interactions.Bioinformatics. 2003 Feb 12;19(3):376-82

Examples

data(mdrdata)
cvk<-10
nbr=3
res<-rmdr.fast(mdrdata,10,3, randomize=TRUE)

[Package Rmdr version 0.1-1 Index]