gausscov (version 0.1.1)

frmch: Robust selection of covariates using Huber's psi-funtion or Hampel's redescending psi-function based on all subsets

Description

Calculates all possible subsets and selects those where each included covariate is significant using a robustified version of flmmdch.R

Usage

frmch(y,x,cn=1,cnr=c(1,2,4),p0=0.01,q=-1,sg=0,ind=0,sel=T,inr=T,xinr=F,red=F)

Arguments

y

Dependent variable

x

Covariates

cn

Constant for Huber's psi-function

cnr

Constants for for Hampel's three part redescending psi-function

p0

The P-value cut-off

q

The numer of covariates available

sg

The scale parameter

ind

The subset for which the results are required

sel

Logical, if TRUE remove all subsets of chosen sets

inr

Logical if TRUE include intercept

xinr

Logical If TRUE intercept already included

red

Logical If true Hampel's three part redescending psi function

Value

nv List of subsets with number of covariates and scale.

Examples

# NOT RUN {
data(boston)
a<-frmch(boston[,14],boston[,1:6])
ind<-decode(57,6)
# }