README.md

In TimothyHyndman/deconvolve: Deconvolution Tools for Measurement Error Problems

deconvolve

The R package deconvolve provides tools for performing non-parametric deconvolution on measurement error problems. It contains functions for finding bandwidths, deconvolved densities and non-parametric regression estimates.

Installation

You can install the development version from Github.

# install.packages("devtools")
devtools::install_github("timothyhyndman/deconvolve")

Usage

library(deconvolve)
W <- (framingham$SBP21 + framingham$SBP22)/2
d <- deconvolve(W)

Available methods

This is a list of every method that this package contains

Deconvolution

• Homoscedastic errors estimated from replicates
• Heteroscedastic errors estimated from replicates
• Errors assumed symmetric when no errors are provided
• Homoscedastic errors supplied
• Heteroscedastic errors supplied

Bandwidth selection

• Plug-in estimator for homoscedastic errors estimated from replicates
• Plug-in estimator for errors assumed symmetric when no errors are provided
• Plug-in estimator for homoscedastic errors
• Plug-in estimator for heteroscedastic errors
• CV for homoscedastic errors
• SIMEX for homoscedastic errors
• SIMEX for homoscedastic errors estimated from replicates

Regression

• Homoscedastic errors estimated from replicates
• Homoscedastic errors supplied

License

This package is free and open source software, licensed under GPL (>=2).

TimothyHyndman/deconvolve documentation built on May 13, 2019, 11:51 p.m.