Split Normal Distribution

Basic Concepts

The split normal distribution (aka the two-piece normal distribution) splices two normal distributions with the same mode but possibly different variances together at the mode.

The probability density function (pdf) of this distribution, denoted N(μ, σ12, σ22), is

Split normal pdf

An alternative expression for the pdf is

Split normal pdf (alternative)

where φ(z) = the pdf of the standard normal distribution at z. Using Excel, we see that φ(z) = NORM.S.DIST(z,FALSE) and φ((x-μ)/σ) = NORM.DIST(x, μ, σ, FALSE).

The cumulative distribution function (cdf) is

Split normal cdf

where Φ(z) = the cdf of the standard normal distribution at z. Using Excel, we see that Φ(z) = NORM.S.DIST(z,TRUE) and Φ((x-μ)/σ) = NORM.DIST(x, μ, σ, TRUE).

The inverse of the cdf is

Split normal inverse cdf

where σ = (σ12)/2.

Properties

Key properties of the split normal distribution are

Split normal distribution properties

Worksheet Functions

Real Statistics Functions: The Real Statistics Resource Pack provides the following functions.

SNORM_DIST(x, μ, σ1, σ2, cum) = the probability density function value f(x) for the split normal distribution when cum = FALSE and the corresponding cumulative distribution function F(x) when cum = TRUE.

SNORM_INV(p, μ, σ1, σ2) = the value x such that SNORM_DIST(x, μ, σ1, σ2, TRUE) = p, i.e. inverse of SNORM_DIST(x, μ, σ1, σ2, TRUE).

References

Julio, J. M. (2007) The fan chart: the technical details of the new implementation
https://www.banrep.gov.co/docum/ftp/borra468.pdf

Wikipedia (2022) Split normal distribution
https://en.wikipedia.org/wiki/Split_normal_distribution

2 thoughts on “Split Normal Distribution”

Leave a Comment