The RATS Software Forum

Hi Tom,

Support I have a model

Measurement: du(t) = beta0 + beta1(t)*dy(t) + e(t)
State: beta1(t) = beta1(t-1) + v(t)

I want to get the series v(t) after estimation. DLM instruction does not have a direct way to get it. So, is it correct if I get the state series beta(t) then compute

set vt = beta1 - beta1{1} (suppose I've got beta1(t) from the xstates vector)

Thanks and regards,

The WHAT option (read W-hat) gives you the smoothed or simulated version if you do TYPE=SMOOTH (which has everything conditional on the full data set) or SIMULATE or CSIMULATE.

I see it. Many thanks Tom.