bde {bde}

Generic bounded density constructor

Description

Function to access all the methods

Usage

bde(dataPoints,dataPointsCache=NULL,estimator,b=length(sample)^{-2/5}, 
    lower.limit=0, upper.limit=1,options=NULL)

Arguments

dataPoints

Vector containing the points to be used to estimate the density.

dataPointsCache

Points where the density has to be estimated. If omitted, 101 points equally distributed in the [lower.limit,upper.limit] interval are used

estimator

Density estimator to be used. This has to be one of the following:

• "betakernel": Chen's beta kernel density estimator

• "vitale": Vitale's Bernstein polynomial based estimator

• "boundarykernel": Boundary kernel based density estimators, as proposed by Muller et al.

• "kakizawa": Kakizawa's density estimators

b

Bandwidth to be used. Note that in the case of Vitale's estimator the m parameter is set at 1/b

lower.limit

a numeric value for the lower limit of the bounded interval for the data

upper.limit

a numeric value for the upper limit of the bounded interval for the data. That is, the data is with the [lower.limit,upper.limit] interval

options

A list containing the different options available for the estimators:

• betakernel:

  • "modified": a logical value indicating whether the modified kernel has to be used or not. False by default

  • "normalization": a string: "none", to use the original kernels, "densitywise" to use the macrobeta kernels and "kernelwise" to use the microbeta kernels. If not specified, no normalization is used

  • "mbc": a string indicating the multiplicative bias correction to be used: "none", no correction is used, "jnl" Hirukawa's JNL approach, "ts" Hirukawa's TS approach. If not specified, no correction is used

  • "c": a numeric value between 0 and 1 corresponding to the c parameter in the TS correction (it is only taken into consideration if TS correction is selected). Default value is set to 0.5

• vitale:

  • "biasreduced": a logical value. If true, Leblanc's bias reduced estimator is used; otherwise the original estimator is used. False by default

• boundarykernel:

  • "mu": numeric parameter to indicate the kind of kernel. Options are 0, for the rectangular function, 1 for Epanechnikov's kernel, 2 for the quadratic and 3 for the biquadratic. Default value is set at 1

  • "method": a string indicating the functions to be used: "Muller94" (default value), "Muller91", "Normalize" or "None"

  • "corrected": a logical value indicating whether Jones' non-negativity correction should be used. By default it is set to false

• kakizawa:

  • "method": a string indicating the function to be used "b1", "b2" or "b3" (default value).

  • "estimator": a Bounded Density estimator. See all accepted classes here with getSubclasses("BoundedDensity"). If no estimator is provided, a Muller94BoundaryKernel estimator with default parameters and the same dataPoints as those give for the Kakizawa estimator is used.

  • "gamma": in case that b1 function is used the gamma parameter is required. This parameter takes 0.5 as default value.