Home › Statistics › Probability Distributions › Gamma Distribution

Gamma Distribution

#gamma-distribution #probability-theory | Updated: Oct 24, 2024 | Read time: 12 min | Verified by Aevum Academic Board

The gamma distribution is a two-parameter family of continuous probability distributions widely used in statistics, queuing theory, reliability engineering, and Bayesian inference. It models the time until a specified number of events occur in a Poisson process, making it particularly valuable for analyzing waiting times and inter-arrival intervals.

Denoted as X ~ Γ(k, θ) or Gamma(α, β) depending on the parametrization convention, the distribution is defined for non-negative real numbers and exhibits flexible skewness depending on its shape parameter.

Mathematical Definition

Probability Density Function (PDF)

f(x; k, θ) = \frac{x^{k-1} e^{-x/θ}}{θ^{kk)} for x > 0, k > 0, θ > 0

Where:

  • k (>0) is the shape parameter, controlling the distribution's skewness and peak.
  • θ (>0) is the scale parameter, determining the spread of the distribution.
  • Γ(k) is the gamma function, a generalization of the factorial function: Γ(k) = \int_0^∞ t^{k-1}e^{-t} dt

In the rate parametrization (α, β), where β = 1/θ, the PDF becomes f(x) = \frac{β^α x^{α-1}e^{-βx}}{Γ(α)}.

Key Properties

  • Mean: E[X] = kθ
  • Variance: Var(X) = kθ²
  • Mode: (k - 1)θ for k ≥ 1
  • Skewness: \frac{2}{\sqrt{k}} (right-skewed, approaches normal as k → ∞)
  • Kurtosis: 6/k + 3
  • Moment Generating Function: M(t) = (1 - θt)^{-k} for t < 1/θ
Note: The gamma distribution is closed under convolution. If X₁ ~ Γ(k₁, θ) and X₂ ~ Γ(k₂, θ) are independent, then X₁ + X₂ ~ Γ(k₁+k₂, θ). This property makes it ideal for modeling cumulative waiting times.

Applications

The gamma distribution's mathematical flexibility and interpretability have cemented its role across numerous scientific and industrial domains:

  1. Reliability Engineering: Models time-to-failure for systems with aging components.
  2. Hydrology & Meteorology: Fits rainfall amounts, flood peak discharges, and drought durations.
  3. Insurance & Finance: Models claim sizes, aggregate loss distributions, and waiting times between transactions.
  4. Bayesian Statistics: Serves as the conjugate prior for the rate parameter (λ) of the Poisson distribution and the inverse scale parameter of the exponential distribution.
  5. Queuing Theory: Extends exponential inter-arrival assumptions to more realistic multi-stage service processes.

References & Further Reading

  • Feller, W. (1971). An Introduction to Probability Theory and Its Applications, Vol. II. Wiley.
  • Johnson, N. L., Kotz, S., & Balakrishnan, N. (1994). Continuous Univariate Distributions, Vol. 1. Wiley.
  • Casella, G., & Berger, R. L. (2002). Statistical Inference. Duxbury.
  • Aevum Encyclopedia Editorial Board. (2024). "Bayesian Conjugate Priors: A Practical Guide". Aevum Press.