markov.parms {dse1} | R Documentation |
Construct a Matrix of the Markov Parameters
markov.parms(model, blocks=NULL)
Construct a matrix with partitions [M0|...|Mi] giving the Markov parameters Mi, i+1 = blocks where each Mi is a p by (m+p) matrix, (m is the dimension of the exogeneous series and p is the dimension of endogeneous series) ie. y(t) = e(t) + M [u'(t)|y'(t-1) | u'(t-1)|y'(t-2)]' This requires that models be transformed so that lagged endogeneous variables are inputs. See Mittnik p1190. If blocks=NULL (the default) then at least 3 blocks are generated, and up to n+1, but the series is truncated if the blocks are effectively zero. This will affect the size of the Hankel matrix.
A matrix
S.Mittnik (1989), Multivariate Time Series Analysis With State Space Models, Computers Math Appl. Vol 17, No 8/9, pp1189-1201.
S.Mittnik (1990), Macroeconomic Forecasting Experience With Balance State Space Models, International Journal Of Forecasting, Vol 6, pp337-348.
S.Mittnik (1990), Forecasting With Balanced State Space Representations of Multivariate Distributed Lag Models. J. of Forecasting, Vol.9, 207-218.
if(is.R()) data("eg1.DSE.data.diff", package="dse1") model <- est.VARX.ls(eg1.DSE.data.diff) markov.parms(model)