Control.Monad.Bayes.Interface

Reexports

Definitions

interface MonadSample : (Type -> Type) -> Type

Parameters: m
Constraints: Monad m
Methods:

random : m Double

  Draws a random value from Uniform(0,1)

uniform : Double -> Double -> m Double

  Uniform(min, max)

normal : Double -> Double -> m Double

  Normal(mean, sd)

gamma : Double -> Double -> m Double

  Gamma(shape, scale) -> m Double

beta : Double -> Double -> m Double

  Beta(alpha, beta) -> m Double

bernoulli : Double -> m Bool

  Bernoulli(prob)

binomial : Nat -> Double -> m Nat

  Binomial(num trials, prob of each trial)

categorical : Vect n Double -> m (Fin n)

  Categorical(probs)

logCategorical : Vect n (Log Double) -> m (Fin n)

  Log-categorical(log-probs)

uniformD : Vect (S n) a -> m a

  Uniform-Discrete(values)

dirichlet : Vect n Double -> m (Vect n Double)

  Dirichlet(concentrations)

discreteUniform : Nat -> m Nat

  DiscUniform(range); should return Nat from 0 to (range - 1)

fromPMF : (Nat -> Double) -> m Nat

  Draw from a discrete distribution using the probability mass function and a sequence of draws from Bernoulli.

geometric : Double -> m Nat

  Geometric(prob)

hypergeometric : Nat -> Nat -> Nat -> m Nat

  Hypergeometric(num elements of "type 1", num elements of "type 2", num samples)

poisson : Double -> m Nat

  Poisson(λ)

Implementations:

MonadSample m => MonadSample (MaybeT m)

MonadSample m => MonadSample (ReaderT r m)

MonadSample m => MonadSample (WriterT w m)

MonadSample m => MonadSample (StateT s m)

MonadSample m => MonadSample (RWST r w s m)

random : MonadSample m => m Double

  Draws a random value from Uniform(0,1)

Totality: total
Visibility: public export

uniform : MonadSample m => Double -> Double -> m Double

  Uniform(min, max)

Totality: total
Visibility: public export

normal : MonadSample m => Double -> Double -> m Double

  Normal(mean, sd)

Totality: total
Visibility: public export

gamma : MonadSample m => Double -> Double -> m Double

  Gamma(shape, scale) -> m Double

Totality: total
Visibility: public export

beta : MonadSample m => Double -> Double -> m Double

  Beta(alpha, beta) -> m Double

Totality: total
Visibility: public export

bernoulli : MonadSample m => Double -> m Bool

  Bernoulli(prob)

Totality: total
Visibility: public export

binomial : MonadSample m => Nat -> Double -> m Nat

  Binomial(num trials, prob of each trial)

Totality: total
Visibility: public export

categorical : MonadSample m => Vect n Double -> m (Fin n)

  Categorical(probs)

Totality: total
Visibility: public export

logCategorical : MonadSample m => Vect n (Log Double) -> m (Fin n)

  Log-categorical(log-probs)

Totality: total
Visibility: public export

uniformD : MonadSample m => Vect (S n) a -> m a

  Uniform-Discrete(values)

Totality: total
Visibility: public export

dirichlet : MonadSample m => Vect n Double -> m (Vect n Double)

  Dirichlet(concentrations)

Totality: total
Visibility: public export

discreteUniform : MonadSample m => Nat -> m Nat

  DiscUniform(range); should return Nat from 0 to (range - 1)

Totality: total
Visibility: public export

fromPMF : MonadSample m => (Nat -> Double) -> m Nat

  Draw from a discrete distribution using the probability mass function and a sequence of draws from Bernoulli.

Totality: total
Visibility: public export

geometric : MonadSample m => Double -> m Nat

  Geometric(prob)

Totality: total
Visibility: public export

hypergeometric : MonadSample m => Nat -> Nat -> Nat -> m Nat

  Hypergeometric(num elements of "type 1", num elements of "type 2", num samples)

Totality: total
Visibility: public export

poisson : MonadSample m => Double -> m Nat

  Poisson(λ)

Totality: total
Visibility: public export

interface MonadCond : (Type -> Type) -> Type

Parameters: m
Constraints: Monad m
Methods:

score : Log Double -> m ()

  Record a likelihood. Note: when calling `score (Exp p)`, p must already be in the log-domain.

Implementations:

MonadCond m => MonadCond (MaybeT m)

MonadCond m => MonadCond (ReaderT r m)

MonadCond m => MonadCond (WriterT w m)

MonadCond m => MonadCond (StateT s m)

MonadCond m => MonadCond (RWST r w s m)

score : MonadCond m => Log Double -> m ()

  Record a likelihood. Note: when calling `score (Exp p)`, p must already be in the log-domain.

Totality: total
Visibility: public export

condition : MonadCond m => Bool -> m ()

Totality: total
Visibility: export

interface MonadInfer : (Type -> Type) -> Type

Parameters: m
Constraints: MonadSample m, MonadCond m
Implementations:

MonadInfer m => MonadInfer (MaybeT m)

MonadInfer m => MonadInfer (ReaderT r m)

MonadInfer m => MonadInfer (WriterT w m)

MonadInfer m => MonadInfer (StateT s m)

MonadInfer m => MonadInfer (RWST r w s m)