ncem.estimators.EstimatorLinear.init_model

EstimatorLinear.init_model(optimizer='adam', learning_rate=0.0001, n_eval_nodes_per_graph=32, l2_coef=0.0, l1_coef=0.0, use_source_type=True, use_domain=False, scale_node_size=False, output_layer='linear', **kwargs)

Initialize a ModelInteractions object.

Parameters

• optimizer (str) – Optimizer.

• learning_rate (float) – Learning rate.

• l2_coef (float) – l2 regularization coefficient.

• l1_coef (float) – l1 regularization coefficient.

• n_eval_nodes_per_graph (int) – Number of nodes per graph.

• use_source_type (bool) – Whether to use source type.

• use_domain (bool) – Whether to use domain information.

• scale_node_size (bool) – Whether to scale output layer by node sizes.

• output_layer (str) – Output layer.

• kwargs – Arbitrary keyword arguments.