[R] heteroscedasticity analysis

[Pfaff, Bernhard]

Have a look at the 'lmtest'-contributed package.

maybe the following function is of use, too. Pls. check, if I got the
calculation of the test statistic right.

arch<-function(x, lag=4)
	{
		xfit<-lm(x)
		res<-xfit$residuals
		ressq<-res^2
#
		tslag<-function(x,d=1)
			{
			 n<-length(x)
    			 c(rep(NA,d),x)[1:n]
			} 
#
		ressq.data<-array(NA,c(length(ressq), lag))
		labels<-paste("ressq.l",1:lag, sep="")
		colnames(ressq.data)<-labels
		for ( i in 1:lag)
			{
			 ressq.data[,i]<-tslag(ressq,d=i)
			}
		residssq<-as.data.frame(ressq.data)
		attach(residssq)
		aux<-as.formula(paste("ressq~",paste(labels,collapse="+")))
		auxfit<-lm(aux)
		auxsum<-summary(auxfit)
		stat<-length(auxsum$residuals)* auxsum$r.squared
		pval<-1-pchisq(stat,df=lag)
		detach(residssq)
		ARCH.Stat<-c(stat,pval)
		ARCH.Stat
	}

As input you have tp provide your model formula, the test itself is based
upon the lm-residuals.

HTH,
Bernhard

-----Original Message-----
From: Vincent Spiesser [mailto:Vincent.Spiesser at univ-tlse1.fr]
Sent: 09 December 2002 15:07
To: r-help at stat.math.ethz.ch
Subject: [R] heteroscedasticity analysis


Hello,
First, sorry for my poor english, I will try to be understood.
It's the first time I try this "r-help mailing list" and I hope it will be 
a success.

I am working on heteroscedasticity analysis. I would like to get the 
"Box-Ljung" and the "Lagrange multipliers" test.
I found the first one in the library "ts", but I can't find the second one. 
Does anybody know how this test can be called.

Vincent Spiesser

______________________________________________
R-help at stat.math.ethz.ch mailing list
http://www.stat.math.ethz.ch/mailman/listinfo/r-help


----------------------------------------------------------------------
If you have received this e-mail in error or wish to read our e-mail 
disclaimer statement and monitoring policy, please refer to 
http://www.drkw.com/disc/email/ or contact the sender.
----------------------------------------------------------------------