model.comp: Model comparison for asreml.

In yzhlinscau/AAfun: ASReml-R Added functions

Description

This function would compare models with different random structure under the same fixed factors.

Usage

model.comp(m1=NULL,m2=NULL,Nml=NULL,mulM=NULL,LRT=NULL,rdDF=NULL)

Arguments

m1	Model 1 asreml results.
m2	Model 2 asreml results.
Nml	A vector with more than 2 asreml results.
mulM	mulM is "TRUE" for a vector's comparison with multi-models, default (FALSE).
LRT	Value TRUE for Likelihood ratio test (LRT), default (FALSE) for no LRT.
rdDF	Value TRUE to control Ddf minus 0.5, default FALSE for LRT with Ddf. If TRUE, Ddf would minus 0.5, ie, Ddf=Ddf-0.5 .

Value

m1	An asreml result,such as "m1.asr"
Nml	A vector, such as "c(m1,m2,m3,m4)"

Author(s)

Yuanzhen Lin <[e-mail]>

References

Yuanzhen Lin. R & ASReml-R Statistics. China Forestry Publishing House. 2016

See Also

Website for instant update: yzhlin-asreml.ys168.com

Examples

library(asreml)
library(AAfun)
df<-PrSpa

fm<-asreml(h5~ 1+Rep,random=~ Fam, 
            subset=Spacing=='3',data=df,maxit=40)
            
fm1a<-asreml(cbind(dj,h5)~ trait+trait:Rep, 
            random=~ us(trait):Fam, rcov=~units:us(trait),
            subset=Spacing=='3',data=df,maxit=40)
fm1b<-asreml(cbind(dj,h5)~ trait+trait:Rep, 
            random=~ diag(trait):Fam, rcov=~units:us(trait),
            subset=Spacing=='3',data=df,maxit=40)

fm3a<-asreml(cbind(dj,h3,h5)~ trait+trait:Rep, 
            random=~ diag(trait):Fam, rcov=~units:us(trait),
            subset=Spacing=='3',data=df,maxit=40)
            
fm3b<-asreml(cbind(dj,h3,h5)~ trait+trait:Rep, 
            random=~ diag(trait):Fam, rcov=~units:us(trait),
            subset=Spacing=='3',data=df,maxit=40)

#####   model comparison    #####
model.comp(m1=fm1a,m2=fm1b)
model.comp(m1=fm1a,m2=fm1b,LRT=TRUE)
model.comp(m1=fm1a,m2=fm1b,LRT=TRUE,rdDF=TRUE)

model.comp(Nml=c(fm3a,fm3b,fm1a,fm1b,fm),mulM=TRUE)
model.comp(Nml=c(fm3a,fm3b,fm1a,fm1b,fm),mulM=TRUE,LRT=TRUE)

yzhlinscau/AAfun documentation built on May 21, 2020, 2:19 p.m.