Class GaussianPrior

Isotropic Gaussian prior: P(theta) = Σ (theta_i - mean)² / (2 var).

Hierarchy (View Summary)

Prior
- GaussianPrior

Index

Constructors

Properties

Methods

Constructors

constructor

new GaussianPrior(mean?: number, variance?: number): GaussianPrior
Parameters
- mean: number = 0
- variance: number = 1
Returns GaussianPrior
Overrides Prior.constructor
- Defined in rank/priors.ts:44

Properties

`Readonly`mean

mean: number

`Readonly`var

var: number

Methods

penalty

penalty(theta: readonly number[]): number
Evaluate the prior penalty P(theta) (added to a negative log-likelihood).
Parameters
- theta: readonly number[]
Returns number
Overrides Prior.penalty
- Defined in rank/priors.ts:53