| Title: | Bandwidth Selection for Level Sets and HDR Estimation |
| Version: | 0.1 |
| Description: | Bandwidth selection for kernel density estimators of 2-d level sets and highest density regions. It applies a plug-in strategy to estimate the asymptotic risk function and minimize to get the optimal bandwidth matrix. See Doss and Weng (2018) <doi:10.48550/arXiv.1806.00731> for more detail. |
| Depends: | R (≥ 3.4.0) |
| License: | GPL-3 |
| Encoding: | UTF-8 |
| LazyData: | true |
| Imports: | ks, numDeriv, Matrix |
| URL: | http://arxiv.org/abs/1806.00731 |
| RoxygenNote: | 6.1.0 |
| NeedsCompilation: | no |
| Packaged: | 2018-10-17 20:36:40 UTC; wgw |
| Author: | Guangwei Weng [aut, cre] |
| Maintainer: | Guangwei Weng <wengx076@umn.edu> |
| Repository: | CRAN |
| Date/Publication: | 2018-10-26 15:30:03 UTC |
Calculate the optiaml bandwidth matrix for highest density region estimation
Description
This function allow you to compute the optiaml bandwidth matrix for highest density region estimation by using a plug-in strategy.
Usage
hdrbs(X, tau, xrange, yrange, gridwidth, init = NULL, maxit = 200,
tol = 1e-06, print_obj = FALSE)
Arguments
X |
a matrix with two columns containing the data from the density function. |
tau |
a probability value between 0 and 1 |
xrange |
a vector of of length 2, e.g., |
yrange |
a vector of of length 2, e.g., |
gridwidth |
width between grid points. |
init |
starting value of the bandwidth matrix for
optimization. If not specified, use direct-plug estimator from
|
maxit |
maximum number of iterations for optimization |
tol |
tolerance value for stopping the optimization algorithm |
print_obj |
a flag (boolean type) indicates printing the loss function values during optimizatin or not. |
Value
the optimal bandwidth matrix.
References
Doss, C.R. and Weng, G., 2018. Bandwidth selection for kernel density estimators of multivariate level sets and highest density regions. arXiv preprint arXiv:1806.00731.
Examples
X <- matrix(rnorm(100),ncol=2)
xrange <- c(-2.5,2.5)
yrange <- c(-2.5,2.5)
hdrbs(X,0.1,xrange,yrange,0.1)
Calculate the optiaml bandwidth matrix for level set estimation
Description
This function allow you to compute the optiaml bandwidth matrix for level set estimation by using a plug-in strategy.
Usage
lsbs(X, levelc, xrange, yrange, gridwidth, init = NULL, maxit = 200,
tol = 1e-06, print_obj = FALSE)
Arguments
X |
a matrix with two columns containing the data from the density function. |
levelc |
a positive value indicating the height of the level set |
xrange |
a vector of of length 2, e.g., |
yrange |
a vector of of length 2, e.g., |
gridwidth |
width between grid points. |
init |
starting value of the bandwidth matrix for
optimization. If not specified, use direct-plug estimator from
|
maxit |
maximum number of iterations for optimization |
tol |
tolerance value for stopping the optimization algorithm |
print_obj |
a flag (boolean type) indicates printing the loss function values during optimizatin or not. |
Value
the optimal bandwidth matrix.
References
Doss, C.R. and Weng, G., 2018. Bandwidth selection for kernel density estimators of multivariate level sets and highest density regions. arXiv preprint arXiv:1806.00731.
Examples
X <- matrix(rnorm(100),ncol=2)
xrange <- c(-3,3)
yrange <- c(-3,3)
lsbs(X,0.1,xrange,yrange,0.05)