The SSM procedure provides state space modeling of univariate and multivariate time series and longitudinal data. State space models encompass an alternative general formulation of multivariate ARIMA models. The SSM procedure includes the following features:
general linear state space models (SSMs)
expressive language to specify an SSM, including flexible and intuitive specification of transition and covariance matrices
easy specification of commonly used SSMs by using only a few keywords
restricted maximum likelihood estimation computed using the (diffuse) Kalman filter algorithm
forecasts, residuals, and full-sample estimations of any linear combination of state variables
residual diagnostics plots
plots for detecting structural breaks