Class Network

Represents a Bayesian Network, or a Dynamic Bayesian Network. To perform inference with a network see the BayesServer.Inference namespace.

Inheritance

object

Network

Implements

INotifyPropertyChanged

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.ReferenceEquals(object, object)

object.ToString()

Namespace: BayesServer

Assembly: BayesServer.dll

Syntax

public sealed class Network : INotifyPropertyChanged

Remarks

A Bayesian network consists of nodes which contain one or more variables. Each node can be connected with a directed link (A link in a Dynamic Bayesian Network can have an associated order which links nodes in different time slices).

<p>
Each node in a Bayesian network must have a distribution specified over its own variables conditional on its parents variables.  For example if a network contains
nodes A and B each with a single variable of the same name, and A is a parent of B, then the distribution at A equals P(A), since A has no parents,
and the distribution at B equals P(B|A).
</p>
<p>Once the nodes, links, and distributions in a Bayesian network have been specified, one or more inference engines
can be created based on the network, to perform probabilistic inference.  See <xref href="BayesServer.Inference.IInference" data-throw-if-not-resolved="false"></xref>.</p>

Constructors

Network()

Initializes a new instance of the Network class.

Declaration

public Network()

Network(string)

Initializes a new instance of the Network class with the specified name.

Declaration

public Network(string name)

Parameters

Type	Name	Description
string	name	The optional name of the network. Can be null.

Properties

CustomProperties

Gets custom properties associated with this instance.

Declaration

public CustomPropertyCollection CustomProperties { get; }

Property Value

Type	Description
CustomPropertyCollection

Remarks

Custom properties allow storage of custom information, which will be saved with the network.

Description

An optional description for the Bayesian network.

Declaration

public string Description { get; set; }

Property Value

Type	Description
string

HasChanged

Indicates whether the network has changed.

Declaration

public bool HasChanged { get; }

Property Value

Type	Description
bool

Remarks

This property is automatically set to false when the network is saved, or loaded.

IsInternal

For internal use only. May be removed in future versions.

Declaration

public bool IsInternal { get; set; }

Property Value

Type	Description
bool

Links

The collection of links in the Bayesian network.

Declaration

[TypeConverter(typeof(ExpandableObjectConverter))]
public NetworkLinkCollection Links { get; }

Property Value

Type	Description
NetworkLinkCollection

Remarks

Note that if a Node is removed from the network, all of its links are automatically removed.

See Also

Link

Name

An optional name for the Bayesian network.

Declaration

public string Name { get; set; }

Property Value

Type	Description
string

NodeGroups

Gets groups which nodes can belong to.

Declaration

public NetworkNodeGroupCollection NodeGroups { get; }

Property Value

Type	Description
NetworkNodeGroupCollection

Remarks

Each node can belong to one or more groups.

Nodes

The collection of nodes in the Bayesian network.

Declaration

[TypeConverter(typeof(ExpandableObjectConverter))]
public NetworkNodeCollection Nodes { get; }

Property Value

Type	Description
NetworkNodeCollection

See Also

Node

SyncNodeVariableName

When true synchronizes Variable names with their containing Node.

Declaration

public static bool SyncNodeVariableName { get; set; }

Property Value

Type	Description
bool

Remarks

The synchronization is one way from the Node to the Variable only.

Synchronization only occurs when a node has a single variable, whose name matches its containing node.

Variables

The collection of variables in the Bayesian network.

Declaration

[TypeConverter(typeof(ExpandableObjectConverter))]
public NetworkVariableCollection Variables { get; }

Property Value

Type	Description
NetworkVariableCollection

See Also

Variable

Methods

Copy()

Makes a copy of the network.

Declaration

public Network Copy()

Returns

Type	Description
Network	The copy.

IsDag()

Determines whether this instance is a Directed Acyclic Graph (DAG) which is a requirement for Bayesian networks.

Declaration

public bool IsDag()

Returns

Type	Description
bool	true if this instance is a DAG; otherwise, false.

IsTree()

Determines whether this instance is a tree (singly connected).

Declaration

public bool IsTree()

Returns

Type	Description
bool	true if this instance is a tree; otherwise, false.

Load(Stream)

Loads a Network from the specified input Stream.

Declaration

public void Load(Stream input)

Parameters

Type	Name	Description
Stream	input	The Stream to load from.

Load(string)

Loads a Network from the specified path.

Declaration

public void Load(string path)

Parameters

Type	Name	Description
string	path	The path and file name to load from.

LoadFromString(string)

Loads a network from a string using UTF-8 encoding.

Declaration

public void LoadFromString(string value)

Parameters

Type	Name	Description
string	value	The persisted string value of the network.

LoadFromString(string, Encoding)

Loads a network from a string using the specified encoding.

Declaration

public void LoadFromString(string value, Encoding encoding)

Parameters

Type	Name	Description
string	value	The persisted string value of the network.
Encoding	encoding	The encoding the network was originally saved in.

Save(Stream)

Saves this Network to the specified output Stream.

Declaration

public void Save(Stream output)

Parameters

Type	Name	Description
Stream	output	The Stream to save to.

Remarks

When saving to file, .bayes is the recommended file extension. The encoding used is UTF-8.

Save(string)

Saves this Network to the specified path overwriting the file if it already exists.

Declaration

public void Save(string path)

Parameters

Type	Name	Description
string	path	The path and name of the file to save to.

Remarks

The recommended file extension is .bayes

SaveToString()

Saves the network to a string, with UTF-8 encoding. For large networks consider saving to file or a stream.

Declaration

public string SaveToString()

Returns

Type	Description
string	The network persisted to a string.

SaveToString(Encoding)

Saves the network to a string, with the specified encoding. For large networks consider saving to file or a stream.

Declaration

public string SaveToString(Encoding encoding)

Parameters

Type	Name	Description
Encoding	encoding	The encoding to use.

Returns

Type	Description
string	The network persisted to a string.

Validate(ValidationOptions)

Validates that the Bayesian network is correctly specified.

Declaration

public void Validate(ValidationOptions options)

Parameters

Type	Name	Description
ValidationOptions	options	Specifies options to control the validation process.

Exceptions

Type	Condition
InvalidNetworkException	Raised when the network is not correctly specified, containing the error message.

ValidateTrialSession()

Evaluation version only.

Declaration

[Browsable(false)]
public static void ValidateTrialSession()

Events

PropertyChanged

Declaration

public event PropertyChangedEventHandler PropertyChanged

Event Type

Type	Description
PropertyChangedEventHandler

Implements

INotifyPropertyChanged

See Also

Node

Link