dear Tom:
In local level model with seasonal, the measurement equation is: yt=ut+rt+et. While calculating measuring error or irregular component is et = y- level, why does not et = y- level -seasonal. I found that in commandeur& koopman(2007) chart4_inflation and chart4_UK_data
hardmann
irregular component of seasonal model?
-
TomDoan
- Posts: 7825
- Joined: Wed Nov 01, 2006 4:36 pm
Re: irregular component of seasonal model?
You're correct. Those should both have:hardmann wrote:dear Tom:
In local level model with seasonal, the measurement equation is: yt=ut+rt+et. While calculating measuring error or irregular component is et = y- level, why does not et = y- level -seasonal. I found that in commandeur& koopman(2007) chart4_inflation and chart4_UK_data
hardmann
set level = xstates(t)(1)
set seasonal = xstates(t)(2)
set fitted = %dot(c,xstates)
set irreg = ukinflation-fitted