TOPICS
Search

Generalized Inverse Gaussian Distribution


The generalized inverse Gaussian distribution GIG(nu,a,b) is the continuous distribution on (0,infty) with probability density function

 p(x)=((a/b)^(nu/2))/(2K_nu(sqrt(ab)))x^(nu-1)e^(-(ax+b/x)/2),

where a,b>0, nu is real, and K_nu(x) is the modified Bessel function of the second kind (Jørgensen 1982). The case nu=-1/2 is an inverse Gaussian distribution after parameter conversion.

Mao and Tian (2026) proved that, if -1/2<nu<=1/2 and nu^'>=nu, then a GIG(nu^',a,b) random variable has the distribution of X+Y, where X has distribution GIG(nu,a,b) and Y>=0 is independent. Their analytic criterion is that x^(nu-nu^')K_(nu^')(x)/K_nu(x) is a completely monotonic function on (0,infty) in the stated parameter range. The lower endpoint nu=-1/2 is not included in this statement.


See also

Completely Monotonic Function, Inverse Gaussian Distribution, Modified Bessel Function of the Second Kind

Explore with Wolfram|Alpha

References

Jørgensen, B. Statistical Properties of the Generalized Inverse Gaussian Distribution. New York: Springer-Verlag, 1982. https://doi.org/10.1007/978-1-4612-5698-4.Mao, Z.-X. and Tian, J.-F. "Complete Monotonicity of Laplace Transform Ratios: Density Criteria for Self-Decomposability and Additive Decomposition." 18 Nov 2026. https://arxiv.org/abs/2611.24348v1.

Cite this as:

Weisstein, Eric W. "Generalized Inverse Gaussian Distribution." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/GeneralizedInverseGaussianDistribution.html

Subject classifications