Correlation Analysis between Non-stationary Seasonal Time Series

["Jan Verbesselt" <Jan.Verbesselt@agr.kuleuven.ac.be>]

Hi S & R specialists,

For some time I'm searching for a method that models the relationship
between two time series (Bi-variate time series analysis) that;
        *one time series can be used to predict the other time series
        *determines how strong the relationship is between the two time
series

The time series (couples of two ts.) I'm working with show a strong
seasonal variation and when using a classical decomposition or
differencing method the residuals stay auto-correlated. 

*Is there a linear regression technique for time series that accounts
for non-stationarity? The scatter plot between the two time series shows
are strong linear relationship (R²=0.85 but cannot be used due to
auto-correlation is there something else?)

*Which multi-variate technique or model could be used to model a time
series so that one time series can be used to predict/model the other
time series? I?m working with R & S-plus 6.1 and the add-on FinMetrics
1.0.  Are there relevant articles about this topic? 

Much appreciated,
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/
_______________________________________________________________________