You need just something like: dRho=0.9; dLambda=0.1; dCos=dRho*cos(dLambda); dSin=dRho*sin(dLambda); mCyPhi=(dCos~dSin)|(-dSin~dCos); then you can combine it with trend transition matrices and other components with diagcat(). Also SsfCombine() can be of some help in building multivariate models. Matteo Pelagatti Assistant Professor Dept. of Statistics University of Milan-Bicocca Tel. +39 02 64485834 -----Messaggio originale----- Da: Floris van Ruth [mailto:[log in to unmask]] Inviato: venerd́ 9 febbraio 2007 17.56 A: [log in to unmask]; Matteo Pelagatti Oggetto: Re: R: R: Multivariate UC models in OX I expected as much, thank you for the confirmation. The only problem I have with constructng the matrices by hand is how exactly to enter the frequency and persistence parameters of the cyclical components. Perhaps you could advise me on this, thanks again. Floris van Ruth