est_cfar: Estimation of a CFAR Process

In ConvFuncTimeSeries/test_t: Nonlinear time series analysis

Description

Estimation of a CFAR process.

Usage

est_cfar(f, p = 3, df_b = 10, grid = 1000)

Arguments

f	the functional time series.
p	CFAR order.
df_b	the degrees of freedom for natural cubic splines. Default is 10.
grid	the number of gird points used to constrct the functional time series and noise process. Default is 1000.

Value

The function returns a list with components:

phi_coef	estimated spline coefficients for convluaional function(s).
phi_func	estimated convoluational function(s).
rho	estimated rho for O-U process (noise process).
sigma	estimated sigma for O-U process (noise process).

References

Liu, X., Xiao, H., and Chen, R. (2016) Convolutional autoregressive models for functional time series. Journal of Econometrics, 194, 263-282.

Examples

phi_func= function(x)   
{
 	return(dnorm(x,mean=0,sd=0.1))
}
grid=1000
y=g_cfar1(grid,5,phi_func)
f_grid=y$cfar
index=seq(1,grid+1,by=10)
f=f_grid[,index]
est=est_cfar(f,1)
b_grid=seq(-1,1,by=1/grid)
par(mfcol=c(1,1))
c1 <- range(est$phi_func)
plot(b_grid,phi_func(b_grid),type='l',col='red',ylim=c1*1.1)
points(b_grid,est$phi_func,type='l')

ConvFuncTimeSeries/test_t documentation built on May 29, 2019, 1:39 p.m.