Estimates Moments for a Truncated Normal Distribution using 'Stan'.

truncnormbayes

truncnormbayes provides functionality to estimate mean and standard deviation for a truncated normal distribution.

Specifically, this package finds the posterior modes for the mean and standard deviation for a truncated normal distribution with one or two known truncation points. The method used extends Bayesian methods for parameter estimation for a singly truncated normal distribution under the Jeffreys prior [see Zhou, X., Giacometti, R., Fabozzi, F. J., & Tucker, A. H. (2014). Bayesian estimation of truncated data with applications to operational risk measurement. Quantitative Finance, 14(5), 863-888. DOI 10.1080/14697688.2012.752103]. This package additionally allows for a doubly truncated normal distribution.

Installation

You can install the development version of truncnormbayes from GitHub with:

# install.packages("devtools")
devtools::install_github("leonkt/truncnormbayes")

Example

library(truncnormbayes)

# generate data from a truncated normal
x <- truncnorm::rtruncnorm(100, a = -1, b = 2, mean = 0.5, sd = 0.5)

# estimate its paramaeters
trunc_est(x, a = -1, b = 2)