Struct GammaPosterior 

pub struct GammaPosterior {
    pub shape: f64,
    pub rate: f64,
    pub shape0: f64,
    pub rate0: f64,
}

Gamma posterior for the rate λ of an Exponential(λ) likelihood.

With a Gamma(α, β) prior on λ (shape α, rate β), observing exponential data xᵢ gives the conjugate update α → α + n, β → β + Σxᵢ. Accordingly:

• mean = α / β is the posterior mean of the rate λ.
• posterior_mean_of_mean = β / (α − 1) is the posterior mean of the mean 1/λ (since 1/λ ~ Inverse-Gamma(α, β)).

Note the two differ by a reciprocal: for exponential data with sample mean m the rate posterior concentrates near 1/m, while the mean posterior concentrates near m. Callers that want “the average of the observed values” must use posterior_mean_of_mean, not mean. Feeding an arbitrary positive parameter value in as if it were exponential data (as the old learner did) is not a coherent model — tune positive parameters with ThompsonSamplingTuner instead.

Fields

shape: f64

Shape parameter (α)

rate: f64

Rate parameter (β)

shape0: f64

Prior shape (α₀)

rate0: f64

Prior rate (β₀)

Implementations

impl GammaPosterior

pub fn vague() -> Self

Create with a vague prior Gamma(1, 0.01).

pub fn new(shape: f64, rate: f64) -> Self

Create with a custom prior Gamma(shape, rate).

pub fn observe(&mut self, value: f64)

Conjugate update for one Exponential(λ) datum value.

pub fn mean(&self) -> f64

Posterior mean of the rate λ (= α / β).

pub fn posterior_mean_of_mean(&self) -> Option<f64>

Posterior mean of the mean 1/λ (= β / (α − 1), defined for α > 1).

1/λ ~ Inverse-Gamma(α, β) has mean β / (α − 1). This is the quantity to use when you want the posterior estimate of the average of the observed positive values.

pub fn mode(&self) -> Option<f64>

Posterior mode of the rate (for shape ≥ 1)

pub fn variance(&self) -> f64

Posterior variance of the rate

pub fn observations(&self) -> f64

Number of observed exponential data points, α − α₀.

pub fn sample<R: Rng>(&self, rng: &mut R) -> f64

Sample the rate λ from the posterior.

pub fn decay(&mut self, factor: f64)

Apply decay toward the stored prior.

Trait Implementations

impl Clone for GammaPosterior

fn clone(&self) -> GammaPosterior

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for GammaPosterior

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for GammaPosterior

fn default() -> Self

Returns the “default value” for a type.