npnorm: Distance-based non-parametric method with normal component.

In xiangjiexue/npfixedcomp: Non-parametric methods for estimating mixing distribution

Description

Compute the non-parametric mixing distrbution

Usage

npnorm.sq(
  data,
  stdev = NULL,
  mix = NULL,
  order = FALSE,
  maxit = 100,
  tol = 1e-06
)

npnorm.cvm(
  data,
  stdev = NULL,
  mix = NULL,
  order = FALSE,
  maxit = 100,
  tol = 1e-06
)

npnorm.ad(
  data,
  stdev = NULL,
  mix = NULL,
  order = FALSE,
  maxit = 100,
  tol = 1e-06
)

Arguments

data	the vector of observation to compute the mixing distribution
stdev	the standard deviation of the component density
mix	the initial mixing distribution
order	the digit to round down the observations. If FALSE, no binning is performed.
maxit	the maximum iterations allowed.
tol	the tolerence. Stop if the directional derivative is less than tol

Details

This function uses the non-parametric method to compute the mixing distribution based on various losses

npnorm.sq computes the mixing distribution under squared error loss.

npnorm.cvm computes the mixing distribution under cramer-von Mises loss.

npnorm.ad computes the mixing distribution under Anderson-Darling loss.

Value

the nspmix object with ll meaning the loss.

Examples

data = rnorm(100, c(0, 2))
npnorm.sq(data)
npnorm.cvm(data)
npnorm.ad(data)
datalarge = rnorm(1e5, c(0, 2))
npnorm.cvm(datalarge, order = -2)
npnorm.ad(datalarge, order = -2)

xiangjiexue/npfixedcomp documentation built on June 15, 2020, 9:18 a.m.