2

I am looking for some help with estimating Space state model of this form:

$r_{t} = r^{*}_{t} + \pi + \varepsilon_{1}$

$R_{t}= r^{*}_{t} + \alpha + \pi + \varepsilon_{2}$

$r^{*}_{t} = r^{*}_{t-1} + \phi_{1}$

$\alpha_{t} = \mu_{0} + \mu_{1} \alpha_{t-1} + \phi_{2}$

The first two are the observations equations. I am hoping to be able to estimate $r^{*}$ and $\alpha$ from this system of equations. I am currently trying to do this with matlab, but my biggest issue is being able to specify this in a form that is implementable in matlab (or R) . I will be extremely grateful for any suggestions I can get or pointers that may lead me on.

magnaJ
  • 131
  • 1
  • 4
  • 1
    Shiver, I am not sure if I got the MathJax for your equations correct. Particularly, in the equation for $\alpha_{t}$ (1) should "$\mu_{1}\alpha_{t-1}$" be "$\mu_{1} + \alpha_{t-1}$"; and (2) should "+2" belong somewhere else (like in a subscript)? – Alexis Feb 12 '15 at 02:15
  • Wow! thanks, i was struggling with that for a bit. it should be φ_2 – magnaJ Feb 12 '15 at 02:15
  • Cool beans. How about the $\mu_{1}\alpha_{t-1}$... they good, or need a plus sign? – Alexis Feb 12 '15 at 02:19
  • This is already written in SSM form. I don't understand where is the struggle. – Aksakal Feb 12 '15 at 02:38
  • Hi Aksakal, thanks for taking a look. I have edited the question to make it clearer. I am quite new to this topic, I have studied the basics, but still unfamiliar with actual implementation. – magnaJ Feb 12 '15 at 02:42
  • Did you look at the ~KFAS package in R? – rbm Apr 09 '15 at 13:35
  • 2
    "How to implement this in R" would be off-topic, and it is unclear in what sense these equations need to / cannot be implemented. There may be an underlying on-topic question about how to convert this into some specific model form (used in some package), but this should be clarified. – Juho Kokkala Jun 29 '17 at 05:51

0 Answers0