User guide

The pages follow the order of DK Part I: the model (ch. 3), its initialization (ch. 5), filtering and smoothing (ch. 4-6), estimation (ch. 7), and the models of ch. 3 and 8. Section numbers such as “DK §4.3” refer to Durbin and Koopman (2012).

• The state space model
  • The representation
  • Missing observations
  • Dimensions and cost
• Initialization
  • Kinds
  • Blocks
  • Exact and approximate diffuse
  • Diffuse elements as regression coefficients
• Filtering and smoothing
  • The Kalman filter
  • Univariate treatment
  • Diffuse periods
  • Steady state
  • State and disturbance smoothing
  • Other smoothers
• Estimation
  • Maximum likelihood
  • The gradient
  • Standard errors
  • Concentrated scale
  • EM
  • Information criteria
  • Uncertainty from estimation
  • Many series
• Structural and ARIMA models
  • Components
  • Trends
  • Regression and interventions
  • ARIMA
  • Several series
  • Continuous time and splines
• Defining models
  • Declared models
  • Starting from components
  • Python models
  • Fortran models
• Forecasting and simulation
  • Forecasting
  • Simulation
  • The effect of estimating the parameters
• Diagnostic checking
  • Residuals
  • Tests
  • Auxiliary residuals
  • Goodness of fit
• Performance
  • Where the time goes