package sampling
 Alphabetic
 Public
 All
Type Members

class
AnytimeElementSampler
extends ElementSampler with UnweightedSampler with AnytimeProbQuerySampler
Anytime Element sampler.

trait
AnytimeMPESampler
extends AnytimeSampler with AnytimeMPE
Anytime sampling algorithms that compute MPE.

class
AnytimeMetropolisHastings
extends MetropolisHastings with UnweightedSampler with AnytimeProbQuerySampler
Anytime MetropolisHastings sampler.

class
AnytimeMetropolisHastingsAnnealer
extends MetropolisHastingsAnnealer with AnytimeMPESampler
Anytime MetropolisHastings annealer.

trait
AnytimeProbEvidenceSampler
extends AnytimeSampler with AnytimeProbEvidence
Anytime sampling algorithms that compute probability of evidence.

trait
AnytimeProbQuerySampler
extends AnytimeProbQuery with AnytimeSampler
Anytime sampling algorithms that compute conditional probability of query elements.

trait
AnytimeSampler
extends Algorithm with Anytime with Sampler
Anytime sampling algorithms.

trait
BaseProbQuerySampler
[U[_]] extends BaseProbQueryAlgorithm[U]
A base trait for sampling algorithms that compute conditional probabilities of queries, and that use the projection of all the samples of a target variable to calculate the distribution of that variable or the expectation of a function on that variable.
A base trait for sampling algorithms that compute conditional probabilities of queries, and that use the projection of all the samples of a target variable to calculate the distribution of that variable or the expectation of a function on that variable. Generic type U is either
Element
orReference
. 
abstract
class
BaseUnweightedSampler
extends Sampler
Samplers that use samples without weights.

case class
DisjointScheme
(choices: (Double, () ⇒ ProposalScheme)*) extends ProposalScheme with Product with Serializable
A proposal scheme that chooses between different proposal schemes, each with a probability.
A proposal scheme that chooses between different proposal schemes, each with a probability.
 choices
A variable list of pairs of probabilities and proposal schemes.

abstract
class
ElementSampler
extends BaseUnweightedSampler
An abstract class to generates samples from the marginal distribution of an element.

case class
FinalScheme
(element: () ⇒ Element[_]) extends ProposalScheme with Product with Serializable
A proposal scheme that consists of proposing a single element.
A proposal scheme that consists of proposing a single element.
 element
The element to propose
 class ForwardWeighter extends LikelihoodWeighter

abstract
class
Importance
extends WeightedSampler
Importance samplers.

class
LikelihoodWeighter
extends AnyRef
A class that implements sampling via likelihood weighting on a set of elements.

abstract
class
MetropolisHastings
extends BaseUnweightedSampler
MetropolisHastings samplers.

abstract
class
MetropolisHastingsAnnealer
extends MetropolisHastings with MPEAlgorithm
MetropolisHastings based Annealer.

class
OneTimeElementSampler
extends ElementSampler with UnweightedSampler with OneTimeProbQuerySampler
Onetime Element sampler.

trait
OneTimeMPESampler
extends OneTimeSampler with OneTimeMPE
Onetime sampling algorithms that compute probability of evidence.

class
OneTimeMetropolisHastings
extends MetropolisHastings with UnweightedSampler with OneTimeProbQuerySampler
Onetime MetropolisHastings sampler.

class
OneTimeMetropolisHastingsAnnealer
extends MetropolisHastingsAnnealer with OneTimeMPESampler
Onetime MetropolisHastings annealer.

trait
OneTimeProbEvidenceSampler
extends OneTimeSampler with OneTimeProbEvidence
Onetime sampling algorithms that compute probability of evidence.

trait
OneTimeProbQuerySampler
extends ProbQueryAlgorithm with OneTimeSampler with OneTimeProbQuery
Onetime sampling algorithms that compute conditional probability of query elements.

trait
OneTimeSampler
extends Sampler with OneTime
Onetime sampling algorithms.

abstract
class
ProbEvidenceSampler
extends ProbEvidenceAlgorithm with Sampler
Algorithm that computes probability of evidence using forward sampling.
Algorithm that computes probability of evidence using forward sampling. The evidence is specified as NamedEvidence. Only the probability of this evidence is computed. Conditions and constraints that are already on elements are considered part of the definition of the model.

trait
ProbQuerySampler
extends BaseProbQuerySampler[Element]
Sampling algorithms that compute conditional probabilities of queries on elements, and that use the projection of all the samples of a target variable to calculate the distribution of that variable or the expectation of a function on that variable.

sealed abstract
class
ProposalScheme
extends AnyRef
Class that describes proposal schemes used in MetropolisHastings algorithms.

trait
Sampler
extends Algorithm
A trait for sampling algorithms.

class
Schedule
extends AnyRef
The schedule class determines the annealing schedule in the annealer.

case class
SwitchScheme
[T, U](firstElems: () ⇒ (Element[T], Element[U]), rest: Option[ProposalScheme]) extends ProposalScheme with Product with Serializable
Proposes switching the randomness of two elements and optionally continues with another proposal scheme.
Proposes switching the randomness of two elements and optionally continues with another proposal scheme.
 firstElems
The two elements whose randomness is switched.
 rest
The optional proposal scheme.

case class
TypedScheme
[T](first: () ⇒ Element[T], rest: (T) ⇒ Option[ProposalScheme]) extends ProposalScheme with Product with Serializable
A proposal scheme that proposes a first element and optionally continues with another proposal scheme based on the value of the first element.
A proposal scheme that proposes a first element and optionally continues with another proposal scheme based on the value of the first element.
 first
The first element to propose.
 rest
An optional proposal scheme that is invoked conditioned on the value of the first element.

case class
UntypedScheme
(first: () ⇒ Element[_], rest: Option[ProposalScheme]) extends ProposalScheme with Product with Serializable
A proposal scheme that proposes a first element and optionally continues with another proposal scheme.
A proposal scheme that proposes a first element and optionally continues with another proposal scheme.
 first
The first element to propose.
 rest
An optional proposal scheme that is invoked after the first element is proposed.
 trait UnweightedSampler extends BaseUnweightedSampler with ProbQuerySampler with StreamableProbQueryAlgorithm

abstract
class
WeightedSampler
extends ProbQuerySampler with Sampler with StreamableProbQueryAlgorithm
Samplers that use weighted samples.
Value Members
 object ElementSampler

object
EmptyProposalScheme
extends RuntimeException
Exception thrown when attempting to create a proposal scheme with no steps.

object
Forward
A forward sampler that generates a state by generating values for elements, making sure to generate all the arguments of an element before the element.
 object Importance
 object MetropolisHastings
 object MetropolisHastingsAnnealer
 object ProbEvidenceSampler
 object ProposalScheme
 object Schedule