io.improbable.keanu.network

Class ModelComposition

java.lang.Object

io.improbable.keanu.network.ModelComposition

public final class ModelComposition
extends java.lang.Object

Method Summary

Modifier and Type	Method and Description
static java.util.Map<VertexLabel,Vertex>	composeModel(BayesianNetwork bayesianNetwork, java.util.Map<VertexLabel,Vertex> inputVertices, java.util.List<VertexLabel> desiredOutputs) Connects one model within an other, marking it such that it's clear this is a nested model

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

composeModel

public static java.util.Map<VertexLabel,Vertex> composeModel(BayesianNetwork bayesianNetwork,
                                                             java.util.Map<VertexLabel,Vertex> inputVertices,
                                                             java.util.List<VertexLabel> desiredOutputs)

Connects one model within an other, marking it such that it's clear this is a nested model

This function will take a previously constructed BayesNet and do a number of things: - Hook up any Proxy Vertices to the parents specified in the inputVertices - Increment the depth for all "internal" nodes and the BayesNet itself - Pass back any specified output nodes from the Bayesnet (keeping them at the outer depth) Output nodes will be returned unlabelled

Parameters:

bayesianNetwork - The Bayesian Network to compose in to the current model

inputVertices - The mapping from Proxy label to actual input vertex

desiredOutputs - The set of labels we wish to output from the supplied BayesNet

Returns:

A map of Labels to Output Vertices