Hi,
has anyone experience or an example how to setup a state space model for time
varying regression coefficient estimates in R and how to get the filtered
coefficient estimates.
The model looks like
y(t) = a(t)'*x(t)+u(t)
where y(t) and x(t) are the observations at t=1,2,...,T. The coefficients
a(t)'=(a_1(t),a_2(t),..,a_n(t)) follow a random walk
a_i(t)=a_i(t-1)+v_i(t).
The disturbances u(t) and v_i(t) are assumed to be normally distributed.
Have found a finance example from Zivot using the SsfPack from S-Plus and have
tried to get familiar with the sspir and the dse package from R but I have
problems to adapt the examples in those packages.
Thanks for your help.
Regards,
Daphne
---------------------------------
[[alternative HTML version deleted]]