FGN (version 2.0-12)

GetFitFD: Fit FD Time Series Model

Description

Exact maximum likelihood estimation of the parameter d in fractionally-differenced white noise (FD). In this model, alpha=1-2*d.

Usage

GetFitFD(z, MeanZeroQ = FALSE, algorithm=c("emle","wmle"), ciQ=FALSE)

Arguments

z

time series data vector

MeanZeroQ

optional argument, default is MeanZeroQ=FALSE. Set to TRUE if the mean is known to be zero

algorithm

"emle" or "wmle" for exact or Whittle mle

ciQ

TRUE or FALSE according as 95 percent confidence interval computed and plotted

Value

a list with four/five elements:

d

MLE for d

Loglikelihood

value of the maximized loglikelihood

alpha

MLE for alpha

algorithm

either "emle" or "wmle"

ci

95 percent confidence interval for d

See Also

GetFitFGN

Examples

#Example 1
#fit Gaussian White Noise, d=0
z<-rnorm(500, 100, 10)
GetFitFD(z)

#Example 2
#estimate d for NileMin series
data(NileMin)
GetFitFD(NileMin)
GetFitFD(NileMin, algorithm="wmle")