Returns the filtered values for the basic time invariant state-space model; inputs are not allowed.
Kfilter0(num, y, A, mu0, Sigma0, Phi, cQ, cR)
number of observations
data matrix, vector or time series
time-invariant observation matrix
initial state mean vector
initial state covariance matrix
state transition matrix
Cholesky-type decomposition of state error covariance matrix Q -- see details below
Cholesky-type decomposition of observation error covariance matrix R -- see details below
one-step-ahead state prediction
mean square prediction error
filter value of the state
mean square filter error
the negative of the log likelihood
innovation series
innovation covariances
last value of the gain, needed for smoothing
Practically, the script only requires that Q or R may be reconstructed as t(cQ)%*%(cQ)
or t(cR)%*%(cR)
, respectively.
http://www.stat.pitt.edu/stoffer/tsa4/
See also http://www.stat.pitt.edu/stoffer/tsa4/chap6.htm for an explanation of the difference between levels 0, 1, and 2.