FitAR (version 1.94)

GetARMeanMLE: Exact MLE for Mean in AR(p)

Description

Details of this algorithm are given in McLeod and Zhang (2007).

Usage

GetARMeanMLE(z, phi)

Arguments

z

vector of length n containing the time series

phi

vector of AR coefficients

Value

Estimate of mean

References

McLeod, A.I. and Zhang, Y. (2006). Partial autocorrelation parameterization for subset autoregression. Journal of Time Series Analysis, 27, 599-612.

See Also

mean

Examples

#Simulate a time series with mean zero and compute the exact
#mle for mean and compare with sample average.
## Not run:  #save time building package!
#  set.seed(3323)
#  phi<-c(2.7607,-3.8106,2.6535,-0.9238)
#  z<-SimulateGaussianAR(phi,1000)
#  ans1<-mean(z)
#  ans2<-GetARMeanMLE(z,phi)
# # define a direct MLE function
# "DirectGetMeanMLE" <-
# function(z, phi){
#     GInv<-solve(toeplitz(TacvfAR(phi, length(z)-1)))
#     g1<-colSums(GInv)
#     sum(g1*z)/sum(g1)
# }
# ans3<-DirectGetMeanMLE(z,phi)
# ans<-c(ans1,ans2,ans3)
# names(ans)<-c("mean", "GetARMeanMLE","DirectGetMeanMLE")
# ans
# ## End(Not run)