Generalized inverse Gaussian distribution

A three-parameter positive continuous distribution with density proportional to x^{p−1}exp[−(ax+b/x)/2], normalized by a modified Bessel K function and containing inverse Gaussian and gamma-related limits.

Version

v2 · 2026-09-19 · History

Domain-specific #

4696

Domain group

Formal Sciences

Origin domain

Mathematics

Subdomains

Probability and Statistics, Continuous Probability Distributions → Mathematics

Aliases

GIG distribution

Core Idea

The generalized inverse Gaussian (GIG) distribution has density f(x)=(a/b)^{p/2}xexp[−(ax+b/x)/2]/(2K_p(√ab)) for x>0 under its standard parameterization. The exponential penalizes both large x through ax and near-zero x through b/x, while the power term controls shape. Bessel-K identities normalize the density and generate moments; reciprocal transformation maps parameters predictably. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

Not-Too-Big, Not-Too-Tiny Rule

Imagine guessing how long you'll wait for a bus, and the wait is always more than zero. Some math rules say how likely each wait is. The Generalized inverse Gaussian distribution is one of these rules, where a super-tiny wait and a super-huge wait are both made unlikely, and a knob changes the shape in between.

The Two-Sided Squeeze Curve

A probability distribution tells you how likely different numbers are. The Generalized inverse Gaussian distribution only covers positive numbers. Its formula has one part that pushes down the chances of big numbers and another part that pushes down the chances of numbers very close to zero. A third part changes the overall shape. Mathematicians use special functions called Bessel functions to make all the chances add up correctly.

Bessel-Normalized Positive Distribution

The Generalized inverse Gaussian (GIG) distribution is a family of probability distributions on positive numbers x, with three parameters a, b, and p. In its standard form, the density contains exp[−(ax + b/x)/2]. The ax term makes large values unlikely, and the b/x term makes values near zero unlikely. A power x^(p−1) controls the shape. To make the total probability equal 1, the formula divides by a modified Bessel function K_p, and the same function also gives the moments. Taking the reciprocal 1/x of a GIG variable gives another GIG variable with predictably changed parameters. The family is used in mixture models.

 

The Generalized inverse Gaussian (GIG) distribution has density f(x) = (a/b)^{p/2} x^{p−1} exp[−(ax + b/x)/2] / (2K_p(√ab)) for x > 0 in its standard parameterization. Here K_p is the modified Bessel function of the second kind. The exponent balances two penalties: ax suppresses large values and b/x suppresses values near zero, while x^{p−1} adjusts the shape. Bessel-K identities supply the normalizing constant and closed forms for moments. The family is closed under reciprocal transformation, which maps parameters predictably. It is used in mixture constructions. Using it correctly requires stating the support, parameter domain, parameterization, normalization, and limiting-case conventions consistently, since several parameterizations exist.

Scope of Application

Generalized inverse Gaussian distribution belongs to probability and statistics and is useful where the analyst can specify a positive random variable x, real shape parameter p, positive scale parameters a and b under the standard case, and the modified Bessel function K_p, then evaluate support, parameter domain, density parameterization, Bessel normalization, and limiting-case convention are stated consistently. The scope is broad within that domain but bounded by the need for support, parameter domain, density parameterization, Bessel normalization, and limiting-case convention are stated consistently. The entry records a descriptive analytical identity; practical use requires the governing domain's evidence, standards, and safety obligations.

Clarity

The abstraction clarifies a crowded vocabulary by making support, parameter domain, density parameterization, Bessel normalization, and limiting-case convention are stated consistently the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test. A bare label is insufficient because the name Generalized inverse Gaussian distribution can be used for a formal identity, an implementation, or a neighboring result unless carrier and convention are stated.

Manages Complexity

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Generalized inverse Gaussian distribution. Generalized inverse Gaussian distribution compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning

1. Identify the carrier. State what the elements, states, objects, or observations are: a positive random variable x, real shape parameter p, positive scale parameters a and b under the standard case, and the modified Bessel function K_p. Reject examples whose alleged carrier belongs to a different problem.
2. Lock the constitutive rule. Express support, parameter domain, density parameterization, Bessel normalization, and limiting-case convention are stated consistently independently of one notation or implementation.

Knowledge Transfer

Knowledge transfers strongly among subfields of probability and statistics because they reuse a positive random variable x, real shape parameter p, positive scale parameters a and b under the standard case, and the modified Bessel function K_p, The exponential penalizes both large x through ax and near-zero x through b/x, while the power term controls shape.

Relationships to Other Abstractions

Current abstraction Generalized inverse Gaussian distribution Domain-specific

Parents (1) — outgoing typed relations

• Generalized inverse Gaussian distribution is a kind of Function (Mapping) Prime

  The density maps positive values and parameters to normalized probability weight.

Hierarchy path (1) — routes to 1 parentless root

• Generalized inverse Gaussian distribution → is a kind of Function (Mapping)

Neighborhood in Abstraction Space

Generalized inverse Gaussian distribution sits in a crowded region of the domain-specific corpus (28^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Multivariate & Heavy-Tailed Distributions (16 abstractions)

Nearest neighbors

• Normal-exponential-gamma distribution — 0.90
• Sombrero function — 0.89
• Quantile function — 0.89
• Variance-gamma distribution — 0.89
• Lommel polynomial — 0.88

Computed from structural-signature embeddings · 2026-10-25