Sspace
State space object. Estimation and evaluation of state space models using the Kalman filter.
Sspace Declaration
To declare a sspace object, use the sspace keyword, followed by a valid name.
Sspace Method
ml maximum likelihood estimation or filter initialization.
Sspace Views
cellipse Confidence ellipses for coefficient restrictions.
coefcov coefficient covariance matrix.
display display table, graph, or spool in object window.
endog table or graph of actual signal variables.
grads examine the gradients of the log likelihood.
label label information for the state space object.
output table of estimation results.
residcor standardized one-step ahead residual correlation matrix.
residcov standardized one-step ahead residual covariance matrix.
resids one-step ahead actual, fitted, residual graph.
results table of estimation and filter results.
spec text representation of state space specification.
statefinal display the final values of the states or state covariance.
stateinit display the initial values of the states or state covariance.
structure examine coefficient or variance structure of the specification.
wald Wald coefficient restriction test.
Sspace Procs
append add line to the specification.
clearhist clear the contents of the history attribute.
copy creates a copy of the sspace.
forecast perform state and signal forecasting.
makeendog make group containing actual values for signal variables.
makegrads make group containing the gradients of the log likelihood.
makemodel make a model object containing equations in sspace.
makesignals make group containing signal and residual series.
olepush push updates to OLE linked objects in open applications.
setattr set the value of an object attribute.
Sspace Data Members
Scalar Values
@coefcov(i,j) covariance of coefficients i and j.
@coefs(i) coefficient i.
@eqncoef(k) number of estimated coefficients in equation k.
@eqregobs(k) number of observations in signal equation k.
@linecount scalar containing the number of lines in the Sspace object.
@sddep(k) standard deviation of the signal variable in equation k.
@ssr(k) sum-of-squared standardized one-step ahead residuals for equation k.
@stderrs(i) standard error for coefficient i.
@tstats(i) t-statistic value for coefficient i.
Scalar Values (system level data)
@aic Akaike information criterion for the system.
@hq Hannan-Quinn information criterion for the system.
@logl value of the log likelihood function.
@ncoefs total number of estimated coefficients in the system.
@neqns number of equations for observable variables.
@regobs number of observations in the system.
@sc Schwarz information criterion for the system.
@totalobs sum of “@eqregobs” from each equation.
Vectors and Matrices
@coefcov covariance matrix for coefficients of equation.
@coefs coefficient vector.
@final_state matrix of final states.
@final_statecov (sym) covariance matrix of final state covariances.
@init_state matrix of initial states.
@init_statecov (sym) covariance matrix of initial state covariances.
@pvals vector containing the coefficient probability values.
@residcov (sym) covariance matrix of the residuals.
@stderrs vector of standard errors for coefficients.
@tstats vector of t-statistic values for coefficients.
State and Signal Results
The following functions allow you to extract the filter and smoother results for the estimation sample and place them in matrix objects. In some cases, the results overlap those available thorough the sspace procs, while in other cases, the matrix results are the only way to obtain the results.
Note also that since the computations are only for the estimation sample, the one-step-ahead predicted state and state standard error values will not match the final values displayed in the estimation output. The latter are the predicted values for the first out-of-estimation sample period.
@pred_signal matrix or vector of one-step ahead predicted signals.
@pred_signalcov matrix where every row is the @vech of the one-step ahead predicted signal covariance.
@pred_signalse matrix or vector of the standard errors of the one-step ahead predicted signals.
@pred_err matrix or vector of one-step ahead prediction errors.
@pred_errcov matrix where every row is the @vech of the one-step ahead prediction error covariance.
@pred_errcovinv matrix where every row is the @vech of the inverse of the one-step ahead prediction error covariance.
@pred_errse matrix or vector of the standard errors of the one-step ahead prediction errors.
@pred_errstd matrix or vector of standardized one-step ahead prediction errors.
@pred_state matrix or vector of one-step ahead predicted states.
@pred_statecov matrix where each row is the @vech of the one-step ahead predicated state covariance.
@pred_statese matrix or vector of the standard errors of the one-step ahead predicted states.
@pred_stateerr matrix or vector of one-step ahead predicted state errors.
@curr_err matrix or vector of filtered error estimates.
@curr_gain matrix or vector where each row is the @vec of the Kalman gain.
@curr_state matrix or vector of filtered states.
@curr_statecov matrix where every row is the @vech of the filtered state covariance.
@curr_statese matrix or vector of the standard errors of the filtered state estimates.
@sm_signal matrix or vector of smoothed signal estimates.
@sm_signalcov matrix where every row is the @vech of the smoothed signal covariance.
@sm_signalse matrix or vector of the standard errors of the smoothed signals.
@sm_signalerr matrix or vector of smoothed signal error estimates.
@sm_signalerrcov matrix where every row is the @vech of the smoothed signal error covariance.
@sm_signalerrse matrix or vector of the standard errors of the smoothed signal error.
@sm_signalerrstd matrix or vector of the standardized smoothed signal errors.
@sm_state matrix or vector of smoothed states.
@sm_statecov matrix where each row is the @vech of the smoothed state covariances.
@sm_statese matrix or vector of the standard errors of the smoothed state.
@sm_stateerr matrix or vector of the smoothed state errors.
@sm_stateerrcov matrix where each row is the @vech of the smoothed state error covariance.
@sm_stateerrse matrix or vector of the standard errors of the smoothed state errors.
@sm_stateerrstd matrix or vector of the standardized smoothed state errors.
@sm_crosserrcov matrix where each row is the @vec of the smoothed error cross-covariance.
String Values
@attr("arg") string containing the value of the arg attribute, where the argument is specified as a quoted string.
@command full command line form of the state space estimation command. Note this is a combination of @method and @options.
@description string containing the Sspace object’s description (if available).
@detailedtype returns a string with the object type: “SSPACE”.
@displayname returns the Sspace object’s display name. If the Sspace has no display name set, the name is returned.
@line(i) returns a string containing the i-th line of the Sspace object.
@name returns the Sspace’s name.
@options command line form of sspace estimation options.
@remarks string containing the sspace object’s remarks (if available).
@smpl sample used for estimation.
@svector returns an Svector where each element is a line of the Sspace object.
@svectornb same as @svector, with blank lines removed.
@type returns a string with the object type: “SSPACE”.
@updatetime returns a string representation of the time and date at which the Sspace was last updated.
Sspace Examples
The one-step-ahead state values and variances from ss01 may be saved using:
vector ss_state=ss01.@pred_state
matrix ss_statecov=ss01.@pred_statecov