oracle.pgx.api.mllib

Class UnsupervisedGnnExplainer

java.lang.Object

oracle.pgx.api.mllib.GnnExplainer

oracle.pgx.api.mllib.UnsupervisedGnnExplainer

public class UnsupervisedGnnExplainer
extends GnnExplainer

GnnExplainer object for UnsupervisedGraphWiseModel.

Since:

22.2

Method Summary

Modifier and Type	Method and Description
<ID> UnsupervisedGnnExplanation<ID>	inferAndExplain(PgxGraph graph, PgxVertex<ID> vertex) Blocking version of inferAndExplainAsync(PgxGraph, PgxVertex).
<ID> PgxFuture<UnsupervisedGnnExplanation<ID>>	inferAndExplainAsync(PgxGraph graph, PgxVertex<ID> vertex) Performs inference on the specified vertex and generates an explanation that contains scores of how important each property and each vertex in the computation graph is for the prediction.
UnsupervisedGnnExplainer	learningRate(double learningRate) Set the learning rate for the explainer.
UnsupervisedGnnExplainer	marginalize(boolean marginalize) Set whether the explainer loss is marginalized over features.
int	numClusters() Get the number of clusters used for the explainer loss.
UnsupervisedGnnExplainer	numClusters(int numClusters) Set the number of clusters to use in the explainer loss.
UnsupervisedGnnExplainer	numOptimizationSteps(int numOptimizationSteps) Set the number of optimization steps for the explainer.
int	numSamples() Get the number of vertex samples used by the explainer.
UnsupervisedGnnExplainer	numSamples(int numSamples) Set the number of vertex samples to use to optimize the explainer.

Methods inherited from class oracle.pgx.api.mllib.GnnExplainer

learningRate, marginalize, numOptimizationSteps

Methods inherited from class java.lang.Object

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

Method Detail

inferAndExplain

public <ID> UnsupervisedGnnExplanation<ID> inferAndExplain(PgxGraph graph,
                                                           PgxVertex<ID> vertex)

Blocking version of inferAndExplainAsync(PgxGraph, PgxVertex). Performs inference on the specified vertex and generates an explanation that contains scores of how important each property and each vertex in the computation graph is for the prediction.

Parameters:

graph - the input graph

vertex - the vertex

Returns:

explanation containing feature importance and vertex importance.

Since:

22.2

inferAndExplainAsync

public <ID> PgxFuture<UnsupervisedGnnExplanation<ID>> inferAndExplainAsync(PgxGraph graph,
                                                                           PgxVertex<ID> vertex)

Performs inference on the specified vertex and generates an explanation that contains scores of how important each property and each vertex in the computation graph is for the prediction.

Parameters:

graph - the input graph

vertex - the vertex

Returns:

explanation containing feature importance and vertex importance.

Since:

22.2

learningRate

public UnsupervisedGnnExplainer learningRate(double learningRate)

Set the learning rate for the explainer.

Parameters:

learningRate - learning rate

Since:

22.2

marginalize

public UnsupervisedGnnExplainer marginalize(boolean marginalize)

Set whether the explainer loss is marginalized over features. This can help in cases where there are important features that take values close to zero. Without marginalization the explainer can learn to mask such features out even if they are important, marginalization solves this by instead learning a mask for the deviation from the estimated input distribution.

Parameters:

marginalize - true if loss should be marginalized, false otherwise

Since:

22.2

numClusters

public int numClusters()

Get the number of clusters used for the explainer loss.

Returns:

number of clusters

Since:

22.2

numClusters

public UnsupervisedGnnExplainer numClusters(int numClusters)

Set the number of clusters to use in the explainer loss. The unsupervised explainer uses k-means clustering to compute the explainer loss that is optimized. If the approximate number of components in the graph is known, it is a good idea to set the number of clusters to this number.

Parameters:

numClusters - number of clusters

Since:

22.2

numOptimizationSteps

public UnsupervisedGnnExplainer numOptimizationSteps(int numOptimizationSteps)

Set the number of optimization steps for the explainer.

Parameters:

numOptimizationSteps - number of optimization steps

Since:

22.2

numSamples

public int numSamples()

Get the number of vertex samples used by the explainer.

Returns:

number of samples

Since:

22.2

numSamples

public UnsupervisedGnnExplainer numSamples(int numSamples)

Set the number of vertex samples to use to optimize the explainer. For the sake of performance, the explainer computes the loss on this number of randomly sampled vertices. Using more samples will be more accurate but will take longer and use more resources.

Parameters:

numSamples - number of samples

Since:

22.2