Building an Unsupervised Anomaly Detection GraphWise Model Using Partitioned Graphs

17.6.4 Building an Unsupervised Anomaly Detection GraphWise Model Using Partitioned Graphs

You can build an Unsupervised Anomaly Detection GraphWise model using partitioned graphs which have different providers and features.

opg4j> var model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder().
         setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features").
         build()

UnsupervisedAnomalyDetectionGraphWiseModel model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder()
    .setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features")
    .build();

params = dict(vertex_input_property_names=["vertex_provider1_features", "vertex_provider2_features"])
model = analyst.unsupervised_anomaly_detection_graphwise_builder(**params)

It is possible to select which providers you want to train or infer on:

opg4j> var model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder().
         setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features").
         build()

UnsupervisedAnomalyDetectionGraphWiseModel model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder()
    .setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features")
    .build();

params = dict(vertex_input_property_names=["vertex_provider1_features", "vertex_provider2_features"])
model = analyst.unsupervised_anomaly_detection_graphwise_builder(**params)

If you wish to control the flow of the embeddings at each layer, you can enable or disable the connections of interest. By default all the connections are enabled.

opg4j> var convLayerConfig = analyst.graphWiseConvLayerConfigBuilder().
         setNumSampledNeighbors(25).
         useVertexToVertexConnection(true).
         useEdgeToVertexConnection(true).
         useEdgeToEdgeConnection(false).
         useVertexToEdgeConnection(false).
         build()
opg4j> var model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder().
         setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features").
         build()

GraphWiseConvLayerConfig convLayerConfig = analyst.graphWiseConvLayerConfigBuilder()
    .setNumSampledNeighbors(10)
    .useVertexToVertexConnection(true)
    .useEdgeToVertexConnection(true)
    .useEdgeToEdgeConnection(false)
    .useVertexToEdgeConnection(false)
    .build();

UnsupervisedAnomalyDetectionGraphWiseModel model = analyst.unsupervisedAnomalyDetectionGraphWiseModelBuilder()
    .setVertexInputPropertyNames("vertex_provider1_features", "vertex_provider2_features")
    .setConvLayerConfigs(convLayerConfig)
    .build();

conv_layer_config = dict(num_sampled_neighbors=25,
                         activation_fn='tanh',
                         weight_init_scheme='xavier',
                         neighbor_weight_property_name=weightProperty,
                         vertex_to_vertex_connection=True,
                         edge_to_vertex_connection=True,
                         vertex_to_edge_connection=False,
                         edge_to_edge_connection=False)

conv_layer = analyst.graphwise_conv_layer_config(**conv_layer_config)

params = dict(vertex_input_property_names=["vertex_provider1_features", "vertex_provider2_features"],
              conv_layer_config=[conv_layer])

model = analyst.unsupervised_anomaly_detection_graphwise_builder(**params)

Parent topic: Using the Unsupervised Anomaly Detection GraphWise Algorithm (Vertex Embeddings and Anomaly Scores)