dlm-js
    Preparing search index...

    Interface InverseGammaPrior

    Inverse-Gamma(shape, rate) prior specification for a variance parameter.

    Parameterisation: p(σ² | α, β) ∝ (σ²)^{−(α+1)} · exp(−β/σ²) where α = shape, β = rate.

    Common choices:

    • Weakly informative: { shape: 2, rate: 1 } (mode = 1/3)
    • Non-informative (Jeffreys): { shape: 0.001, rate: 0.001 }
    interface InverseGammaPrior {
        rate: number;
        shape: number;
    }

    Properties

    rate: number

    Rate parameter β > 0.

    shape: number

    Shape parameter α > 0.