*
*  Illustration from section 2.10.3
*  Example of estimating free coefficients
*
open data nile.dat
calendar 1871
data(format=free,org=columns,skips=1) 1871:1 1970:1 nile
*
* To do the diffuse KF, use the option "EXACT". To estimate the model with
* sigma-epsilon**2 concentrated out, peg SV to 1 and use the option SCALE.
*
* The log likelihood value in text seems to be omitting the (inessential)
*  -n/2 log(2 pi)-(n-1)/2
*
nonlin psi
dlm(y=nile,method=bfgs,a=1.0,c=1.0,exact,scale,sv=1.0,sw=exp(psi)) 1871:1 1970:1