TBuildList: Construction of a list of regular and irregular histograms,...

Description Usage Arguments Value Author(s) References See Also Examples

Description

Given a sample X, builds a list containing all the desired estimators with specified smoothing parameters.

Usage

1
2
TBuildList(X,family=c("Kernel","RegularHisto","IrregularHisto","Parametric"),
	kerneltab=NULL,bwtab=NULL,Dmax=NULL,Dtab=NULL)

Arguments

X

numeric vector. The sample to build the estimators.

family

vector of estimator types. Default family=c("Kernel", "RegularHisto", "IrregularHisto", "Parametric"). Has to be a subset of the default value.

kerneltab

vector of kernel types, required when the family argument contains 'Kernel'. Should be available types of the generic function 'density'.

If NULL (default) kerneltab="epanechnikov".

bwtab

numeric vector of bandwidths, required when the family argument contains 'Kernel'.

If NULL (default) bwtab = diff(range(X))/2/(ceiling(n/log(n)):1).

Dmax

maximum number of bins, required when the family argument contains 'RegularHisto' or 'IrregularHisto'.

If NULL (default) Dmax=ceiling(n/log(n)).

Dtab

vector of number of bins. See TBuildRegularHisto.

If NULL (default) Dtab=1:Dmax.

Value

TBuildList returns a list containing all constructed estimators, each one consisting in a descriptor list.

Author(s)

Nelo Magalhães and Yves Rozenholc.

References

N. Magalhães and Y. Rozenholc, "A non-combinatorial algorithm for T-estimation", (2014).

See Also

DensityTestim, TBuildRegularHisto, TBuildIrregularHisto, TBuildKernel, TBuildParametric

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
## Not run: 
## list of estimators containing the regular histograms with number of bins varying 
## between 1 and 150, kernel estimators using triangular kernel and parametric estimators 
## build from a normal sample of size 1000:
TBuildList(X=rnorm(1000),family = c("Kernel", "RegularHisto","Parametric"), 
	kerneltab = 'triangular',Dmax=150)

## list of estimators containing irregular histograms and kernel estimators 
## with bandwidths 2^-j, j=1:,...,ceiling(log(length(X))), build from an exponential 
## sample X of size 500:
TBuildList(X=rexp(500),family = c("IrregularHisto","Kernel"),
	bwtab=2^-c(1:ceiling(log(length(X)))))

## End(Not run)

Density.T.HoldOut documentation built on May 2, 2019, 2:32 a.m.