Abstract
State space models have had a tremendous impact on the analysis of time series. Even though the models and ideas are much older the work that Mike West and others started in the 1980ies brought the attention to the statisticians and the models and inferential possibilities have developed enormously. The conceptual idea behind the state space model is that the evolution over time in the object we are observing and the measurement process itself are modelled separately.
My very first serious analysis of a data set was done using a state space model, and since then I seem to have been "haunted" by state space models.
I will not try to give a thorough exposition of the development from simple linear Gaussian state space models to the highly non-linear models analysed with computer intensive methods. Instead I will give examples of some health related applications, that I have been involved in, and relate these to the general development of state space models.
Original language | English |
---|---|
Title of host publication | NBBC07 : Plenary Speakers |
Number of pages | 1 |
Publisher | Research Centre Foulum |
Publication date | 2007 |
Publication status | Published - 2007 |
Event | Nordic-Baltic Biometric Conference - Foulum, Denmark Duration: 6 Jun 2007 → 8 Jun 2007 Conference number: 1 |
Conference
Conference | Nordic-Baltic Biometric Conference |
---|---|
Number | 1 |
Country/Territory | Denmark |
City | Foulum |
Period | 06/06/2007 → 08/06/2007 |