17.3.2 Building a Minimal Supervised EdgeWise Model

You can build an EdgeWise model using the minimal configuration and default hyper-parameters as described in the following code. Note that even though only one feature property is needed (either on vertices with setVertexInputPropertyNames or edges with setEdgeInputPropertyNames) for the model to work, you can specify as many as required.

• JShell
• Java
• Python

opg4j> var model = analyst.supervisedEdgeWiseModelBuilder().
         setVertexInputPropertyNames("vertex_features").
         setEdgeInputPropertyNames("edge_features").
         setEdgeTargetPropertyName("label").
         build()

SupervisedEdgeWiseModel model = analyst.supervisedEdgeWiseModelBuilder()
    .setVertexInputPropertyNames("vertex_features")
    .setEdgeInputPropertyNames("edge_features")
    .setEdgeTargetPropertyName("labels")
    .build();

params = dict(edge_target_property_name="label",
              vertex_input_property_names=["vertex_features"],
              edge_input_property_names=["edge_features"])

model = analyst.supervised_edgewise_builder(**params)