s-news
[Top] [All Lists]

Re: Ljung-Box test in S-Plus

To: rgusso@unive.it, grick@libero.it, s-news <s-news@lists.biostat.wustl.edu>
Subject: Re: Ljung-Box test in S-Plus
From: Patrick Burns <pburns@pburns.seanet.com>
Date: Thu, 20 Mar 2003 17:42:09 +0000
References: <1048172690.783.10.camel@localhost.localdomain>
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.2) Gecko/20021120 Netscape/7.01
I haven't looked specifically at the S-PLUS implementation, but I would think that
it is standard, and this is a lexical question as you say.

Since the test statistic is only based on autocorrelations and the test is quite robust to failures of various assumptions, I tend to think of their statement as appropriate.

There is a working paper on my website that explores the robustness of the test
in some settings.

Patrick Burns

Burns Statistics
patrick@burns-stat.com
+44 (0) 208 525 0696
http://www.burns-stat.com/    (new home of S Poetry)


Riccardo Gusso wrote:

Hello,
the null hypothesis in the Ljung and Box test should be strict white
noise, that is i.i.d., finite variance random variables; but in S-Plus
autocorTest with method "lb" output I read:

Test for Autocorrelation: Ljung-Box

Null Hypothesis: no autocorrelation

Is this only a lexical question, or the S-Plus implementation of Ljung
and Box test is testing only the absence of correlation between random
variables?

Bye,
         Riccardo
--------------------------------------------------------------------
This message was distributed by s-news@lists.biostat.wustl.edu.  To
unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
the BODY of the message:  unsubscribe s-news





<Prev in Thread] Current Thread [Next in Thread>