This chapter surveys state space modelling approaches for analyzing non-normal time series or longitudinal data. The data situation is the same as in Chapter 6, i.e., categorical, counted or nonnegative data are observed over time. Typical examples are categorized daily rainfall data, the number of monthly polio incidences, or daily measurements on sulfur dioxide. State space models, also termed dynamic models, relate time series observations or longitudinal data {y t } to unobserved “states” α t by an observation model for y t given α t . The states, which may be, e.g., unobserved trend and seasonal components or time-varying covariate effects, are assumed to follow a stochastic transition model. Given the observations {y t }, estimation of states (“filtering” and “smoothing”) is a primary goal of inference.


State Space Model Covariate Effect Observation Model Conditional Density Correction Step 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer Science+Business Media New York 1994

Authors and Affiliations

  • Ludwig Fahrmeir
    • 1
  • Gerhard Tutz
    • 2
  1. 1.Seminar für StatistikUniversität MünchenMünchenGermany
  2. 2.Institut für Quantitative MethodenTechnische Universität BerlinBerlinGermany

Personalised recommendations