com.bayesserver

Interface Distribution

All Known Implementing Classes:

CLGaussian, Table

public interface Distribution

Interface specifying the required methods and properties for a probability distribution. For example the Table class implements this interface. While a distribution is assigned to a Node it cannot be modified.

See Also:

Table, CLGaussian

Method Summary

Modifier and Type	Method and Description
Distribution	copy() Creates a copy of the distribution.
Distribution	copy(Integer timeShift) Creates a copy of the distribution, and shifts any times associated with variables by the specified amount.
Distribution	divide(Distribution subset) Creates a new distribution by dividing the instance by the specified subset.
boolean	getLocked() Locks or unlocks a distribution.
Distribution	getOuter() Returns the parent distribution, if this instance is aggregated by another distribution.
Node	getOwner() Gets the current owner, if assigned to a node.
VariableContextCollection	getSortedVariables() Gets the collection of variables in the distribution, sorted by time (which may be null) and the order in which variables were created.
Table	getTable() Gets the Table which specifies the distribution over any discrete variables.
Distribution	instantiate(Double[] values) Calculates the distribution which results from instantiating a number of variables.
boolean	isReadOnly() Indicates whether the distribution is read only.
void	marginalize(Distribution superset) Marginalizes (sums/integrates) the [superset] into this instance.
void	marginalize(Distribution superset, PropagationMethod propagation) Marginalizes (sums/integrates) the [superset] into this instance.
Distribution	multiply(Distribution distribution) Creates a new distribution which is the result of multiplying this instance by another distribution.
void	setLocked(boolean value) Locks or unlocks a distribution.
void	timeShift(int units) Shifts any times associated with the distribution variables by the specified number of time units.

Method Detail

getSortedVariables

VariableContextCollection getSortedVariables()

Gets the collection of variables in the distribution, sorted by time (which may be null) and the order in which variables were created.

Returns:

Variables sorted by time and the order in which variables were created.

timeShift

void timeShift(int units)

Shifts any times associated with the distribution variables by the specified number of time units.

Parameters:

units - The number of time units to shift. Can be negative if required.

getLocked

boolean getLocked()

Locks or unlocks a distribution. The distribution cannot be modified if either getLocked() is true or getOwner() is not null.

Returns:

true if locked; otherwise, false.

setLocked

void setLocked(boolean value)

Locks or unlocks a distribution. The distribution cannot be modified if either getLocked() is true or getOwner() is not null.

Parameters:

value - true if locked; otherwise, false.

isReadOnly

boolean isReadOnly()

Indicates whether the distribution is read only. A distribution is read only if it is locked or is assigned to a Node.

Returns:

true if read only; otherwise, false.

getOwner

Node getOwner()

Gets the current owner, if assigned to a node. A distribution cannot be modified when it is assigned to a node.

getOuter

Distribution getOuter()

Returns the parent distribution, if this instance is aggregated by another distribution.

multiply

Distribution multiply(Distribution distribution)

Creates a new distribution which is the result of multiplying this instance by another distribution.

Parameters:

distribution - The distribution to multiply this instance by.

Returns:

The combined distribution.

divide

Distribution divide(Distribution subset)

Creates a new distribution by dividing the instance by the specified subset. Also known as the complement. The returned distribution, if multiplied by the subset, would be equivalent to this instance.

Parameters:

subset - The subset to divide by.

Returns:

A new distribution.

instantiate

Distribution instantiate(Double[] values)

Calculates the distribution which results from instantiating a number of variables. [values] should contain one entry for each

Parameters:

values - The instantiated values. Entries can be null.

Returns:

The instantiated distribution.

See Also:

in the distribution, however entries can be null.

marginalize

void marginalize(Distribution superset)

Marginalizes (sums/integrates) the [superset] into this instance.

Parameters:

superset - A distribution whose variables form a superset of the variables in this instance.

marginalize

void marginalize(Distribution superset,
                 PropagationMethod propagation)

Marginalizes (sums/integrates) the [superset] into this instance.

Parameters:

superset - A distribution whose variables form a superset of the variables in this instance.

propagation - The propagation method to use during marginalization.

copy

Distribution copy()

Creates a copy of the distribution. The new distribution will not have an owner.

Returns:

A copy of the distribution.

copy

Distribution copy(Integer timeShift)

Creates a copy of the distribution, and shifts any times associated with variables by the specified amount. The new distribution will not have an owner.

Parameters:

timeShift - The amount to shift any times present in the distribution. Can be negative.

Returns:

A copy of this instance, with shifted times.

getTable

Table getTable()

Gets the Table which specifies the distribution over any discrete variables. Often referred to as a Mixture.

Returns:

The table.