s-news
[Top] [All Lists]

Strong Seasonal time Series: STL or VAR with seasonal dummies

To: <s-news@lists.biostat.wustl.edu>
Subject: Strong Seasonal time Series: STL or VAR with seasonal dummies
From: "Jan Verbesselt" <Jan.Verbesselt@agr.kuleuven.ac.be>
Date: Fri, 23 Apr 2004 14:48:16 +0200
Importance: Normal
Hi S helpers,

I'm dealing with strong seasonal bi-variate time series (serie1 and
serie2) and would like to model the relation between both time series.
In S-plus VAR models with extra seasonal dummy variables as exogenous
variables could be the solution. 

==> Does anyone know how I can add this type of variables in the formula
for a VAR model?

By using the "Granger Causality" the aim is then to prove that series1
causes series 2 via the Wald Statistic.

==> What is the definition of R, The restriction matrix, needed for the
Wald Statistic? I don't know how I can create it.

Another solution could be to de-seasonalise the time series via the STL
method and then VAR model the relation and causality. But can the STL
method be inverted so that at the end one can obtain a model which
describes Series2 in function of Series1?  Series2= Seasonal Parameters
+ B* Series1 + Error term (p,q). Or can this be done by an OLS, time
series regression?

Any help, articles, guidelines and experience with "strong seasonal"
time series is mostly welcome!!! 

(I'm working with S-plus 6.1 /Finmetrics1.0 and R 1.9)

Thanks a lot, 
Jan



_______________________________________________________________________
Jan Verbesselt 
Research Associate 
Lab of Geomatics and Forest Engineering K.U. Leuven
Vital Decosterstraat 102. B-3000 Leuven Belgium 
Tel:+32-16-329750   Fax: +32-16-329760
http://gloveg.kuleuven.ac.be/
_______________________________________________________________________



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