Initial condition for Kalman filter

Topics: Kalman Filtering
Apr 22, 2014 at 10:13 AM
Hi,

When using the function 'filter', one can change the initial conditions in the Kalman filter algorithm using the {'initCond','XXX'}. In particular, according to the IRIS user manual, using {'initCond','optimal'} starts the filter from a fixed state vector that is estimated optimally. I am interested in how exactly this initialization procedure works. Can you provide some references to papers (or other kinds of academic literature) that describes this initialization procedure?

Thanks!

Best,

Esben
Coordinator
Apr 22, 2014 at 11:34 AM
Hi Esben,

The initial conditions can be concentrated out of the likelihood and estimated conditional on the parameter vector. See the discussion on Rosenberg's algorithm in Chapter 3 of Harvey's Forecasting: Structural Time Series Models and the Kalman Filter.

Best,

Michael
Marked as answer by jaromirbenes on 5/8/2014 at 9:23 AM
Apr 22, 2014 at 12:20 PM
Thanks, Michael. I will have a look at that piece of literature.

Esben


2014-04-22 13:34 GMT+02:00 michaeljohnston <[email removed]>:

From: michaeljohnston

Hi Esben,

The initial conditions can be concentrated out of the likelihood and estimated conditional on the parameter vector. See the discussion on Rosenberg's algorithm in Chapter 3 of Harvey's Forecasting: Structural Time Series Models and the Kalman Filter.

Best,

Michael

Read the full discussion online.

To add a post to this discussion, reply to this email ([email removed])

To start a new discussion for this project, email [email removed]

You are receiving this email because you subscribed to this discussion on CodePlex. You can unsubscribe on CodePlex.com.

Please note: Images and attachments will be removed from emails. Any posts to this discussion will also be available online at CodePlex.com