h.amise: AMISE Reference (Maximal Smoothing) Bandwidth

View source: R/h.amise.R

h.amiseR Documentation

AMISE Reference (Maximal Smoothing) Bandwidth

Description

Computes the reference bandwidth h_{os} for the r'th derivative of a kernel density estimator: the AMISE-optimal bandwidth in which R(f^{(r+2)}) is replaced by the value given by the maximal smoothing principle of Terrell (1990). For r = 0 it is the oversmoothed bandwidth, an upper bound for the optimal one. Returns an S4 object of class bandwidth.

Usage

h.amise(x, deriv.order = 0L, kernel = "gaussian")

Arguments

x

numeric vector of data values.

deriv.order

non-negative integer: derivative order r. Maximum r for this selector: any for Gaussian and Silverman; cosine 1; logistic 8; Epanechnikov 1; uniform 0; triangular 0; biweight 2; triweight 3; tricube 3; pow(s,t) \min(r^*, 6), where r^* is the largest integer r < t + 1/2 (and r < s + 1/2 if s is not an even integer, and 2r \le st if s and t are integers). Above it h = NA is returned with a warning. See kedd-package, section ‘Maximum derivative order’.

kernel

character string giving the kernel; one of "gaussian" (default), "epanechnikov", "uniform", "triangular", "biweight", "triweight", "tricube", "cosine", "silverman", "logistic", or a power kernel "pow(s,t)" (see kernel.pow).

Details

The AMISE of the r'th derivative estimator is (Scott 1992, p.\,131):

AMISE(h;r)=\frac{R(K^{(r)})}{nh^{2r+1}} + \frac{\mu_2^2(K)}{4}\,h^{4}\,R\!\left(f^{(r+2)}\right)

The minimising bandwidth is:

h_r^{\ast} = \left[\frac{(2r+1)\,R(K^{(r)})}{\mu_2^2(K)\,R(f^{(r+2)})}\right]^{1/(2r+5)} n^{-1/(2r+5)}

R(f^{(r+2)}) is replaced by 35/(243\,\hat{\sigma}^{2r+5}) (maximal smoothing principle, Terrell 1990), where \hat{\sigma} is the sample standard deviation, which gives

h_{os} = \hat{\sigma}\left[\frac{243\,(2r+1)\,R(K^{(r)})}{35\,\mu_2^2(K)\,n}\right]^{1/(2r+5)}.

The same value bounds the search interval of the cross-validation selectors ([0.01\,h_{os}, 2\,h_{os}]).

No optimisation is performed; the result is computed directly. R(K^{(r)})=\int K^{(r)}(x)^2\,dx, \mu_2(K)=\int x^2 K(x)\,dx.

Value

An S4 object of class bandwidth with slots:

data

the (cleaned) data vector.

kernel

name of the kernel used.

deriv.order

derivative order r.

h

reference bandwidth h_{os}.

min.obj

NA (no minimisation is performed).

hos

same as h (the reference bandwidth).

method

"AMISE".

Author(s)

Arsalane Chouaib Guidoum acguidoum@univ-tam.dz, Amel Henider amel.henider@univ-tam.dz and Iago Giné-Vázquez iago.gin-vaz@protonmail.com

References

Radhey, S. S. (1987). MISE of kernel estimates of a density and its derivatives. Statistics and Probability Letters, 5, 153–159.

Scott, D. W. (1992). Multivariate Density Estimation. Theory, Practice and Visualization. New York: Wiley.

Silverman, B. W. (1986). Density Estimation for Statistics and Data Analysis. Chapman & Hall/CRC. London.

Terrell, G. R. (1990). The maximal smoothing principle in density estimation. Journal of the American Statistical Association, 85, 470–477.

Wand, M. P. and Jones, M. C. (1995). Kernel Smoothing. Chapman and Hall, London.

See Also

bandwidth, h.ucv, h.bcv, h.ccv, h.mcv, h.tcv, h.scv, h.mlcv, dkde.

Examples

x <- rnorm(100)

## density (r = 0)
bw0 <- h.amise(x, deriv.order = 0)
show(bw0)

## first derivative (r = 1)
bw1 <- h.amise(x, deriv.order = 1)
show(bw1)
## other kernels on a Marron-Wand sample
h.amise(kurtotic, kernel = "logistic")
h.amise(kurtotic, kernel = kernel.pow(2, 4))

kedd documentation built on Nov. 12, 2026, 5:10 p.m.