8.2.3 Advanced Hyperparameter Customization

You can build a GraphWise model using rich hyperparameter customization.

This is done through the following two sub-config classes:

1. GraphWiseConvLayerConfig
2. GraphWisePredictionLayerConfig

You can create a model with one of the following options:

• only vertex properties
• only edge properties
• both vertex and edge properties

The following code describes the implementation of the configuration using the preceding classes in GraphWise model. The example also specifies a weight decay parameter of 0.001 for the GraphWise model to counteract overfitting.

• JShell
• Java
• Python

opg4j> var weightProperty = analyst.pagerank(trainGraph).getName();
opg4j> var convLayerConfig = analyst.graphWiseConvLayerConfigBuilder().
                setNumSampledNeighbors(25).
                setActivationFunction(ActivationFunction.TANH).
                setWeightInitScheme(WeightInitScheme.XAVIER).
                setWeightedAggregationProperty(weightProperty).
                build()
opg4j> var predictionLayerConfig = analyst.graphWisePredictionLayerConfigBuilder().
                setHiddenDimension(32).
                setActivationFunction(ActivationFunction.RELU).
                setWeightInitScheme(WeightInitScheme.HE).
                build()
opg4j> var model = analyst.supervisedGraphWiseModelBuilder().
                setVertexInputPropertyNames("vertex_features").
                setEdgeInputPropertyNames("edge_features").
                setVertexTargetPropertyName("labels").
                setConvLayerConfigs(convLayerConfig).
                setPredictionLayerConfigs(predictionLayerConfig).
                setWeightDecay(0.001).
                build()

String weightProperty = analyst.pagerank(trainGraph).getName()
GraphWiseConvLayerConfig convLayerConfig = analyst.graphWiseConvLayerConfigBuilder()
    .setNumSampledNeighbors(25)
    .setActivationFunction(ActivationFunction.TANH)
    .setWeightInitScheme(WeightInitScheme.XAVIER)
    .setWeightedAggregationProperty(weightProperty)
    .build();

GraphWisePredictionLayerConfig predictionLayerConfig = analyst.graphWisePredictionLayerConfigBuilder()
    .setHiddenDimension(32)
    .setActivationFunction(ActivationFunction.RELU)
    .setWeightInitScheme(WeightInitScheme.HE)
    .build();

SupervisedGraphWiseModel model = analyst.supervisedGraphWiseModelBuilder()
    .setVertexInputPropertyNames("vertex_features")
    .setEdgeInputPropertyNames("edge_features")
    .setVertexTargetPropertyName("labels")
    .setConvLayerConfigs(convLayerConfig)
    .setPredictionLayerConfigs(predictionLayerConfig)
    .setWeightDecay(0.001)
    .build();

weightProperty = analyst.pagerank(train_graph).name

conv_layer_config = dict(num_sampled_neighbors=25,
                         activation_fn='tanh',
                         weight_init_scheme='xavier',
                         neighbor_weight_property_name=weightProperty)

conv_layer = analyst.graphwise_conv_layer_config(**conv_layer_config)

pred_layer_config = dict(hidden_dim=32,
                         activation_fn='relu',
                         weight_init_scheme='he')

pred_layer = analyst.graphwise_pred_layer_config(**pred_layer_config)

params = dict(vertex_target_property_name="labels",
              conv_layer_config=[conv_layer],
              pred_layer_config=[pred_layer],
              vertex_input_property_names=["vertex_features"],
              edge_input_property_names=["edge_features"],
              seed=17,
              weight_decay=0.001)

model = analyst.supervised_graphwise_builder(**params) 

See SupervisedGraphWiseModelBuilder, GraphWiseConvLayerConfigBuilder and GraphWisePredictionLayerConfigBuilder in Javadoc for a full description of all available hyperparameters and their default values.