Struct GraphNeuralNetwork 

pub struct GraphNeuralNetwork<S = Untrained> { /* private fields */ }

Graph Neural Network for Structured Output Prediction

A graph neural network implementation for multi-output prediction tasks where the outputs have structural relationships represented as a graph. This method can leverage node features, edge information, and graph topology to make predictions that respect the underlying graph structure.

Examples

use sklears_multioutput::{GraphNeuralNetwork, MessagePassingVariant, AggregationFunction};
use sklears_core::traits::{Predict, Fit};
// Use SciRS2-Core for arrays and random number generation (SciRS2 Policy)
use scirs2_core::ndarray::array;

// Node features and adjacency matrix
let node_features = array![[1.0, 2.0], [2.0, 3.0], [3.0, 1.0], [4.0, 4.0], [1.0, 3.0]];
let adjacency = array![[0, 1, 1, 0, 0], [1, 0, 1, 1, 0], [1, 1, 0, 0, 1],
                       [0, 1, 0, 0, 1], [0, 0, 1, 1, 0]];
let node_labels = array![[1, 0, 1], [0, 1, 0], [1, 1, 0], [0, 0, 1], [1, 0, 0]];

let gnn = GraphNeuralNetwork::new()
    .hidden_dim(16)
    .num_layers(2)
    .message_passing_variant(MessagePassingVariant::GCN)
    .aggregation_function(AggregationFunction::Mean);
let trained_gnn = gnn.fit_graph(&adjacency.view(), &node_features.view(), &node_labels).unwrap();
let predictions = trained_gnn.predict_graph(&adjacency.view(), &node_features.view()).unwrap();

Implementations

impl GraphNeuralNetwork<Untrained>

pub fn new() -> Self

Create a new GraphNeuralNetwork instance

pub fn hidden_dim(self, hidden_dim: usize) -> Self

Set the hidden dimension

pub fn num_layers(self, num_layers: usize) -> Self

Set the number of layers

pub fn message_passing_variant(self, variant: MessagePassingVariant) -> Self

Set the message passing variant

pub fn aggregation_function(self, function: AggregationFunction) -> Self

Set the aggregation function

pub fn learning_rate(self, learning_rate: Float) -> Self

Set the learning rate

pub fn max_iter(self, max_iter: usize) -> Self

Set the maximum number of iterations

pub fn dropout_rate(self, dropout_rate: Float) -> Self

Set the dropout rate

pub fn random_state(self, random_state: u64) -> Self

Set random state for reproducible results

impl GraphNeuralNetwork<Untrained>

Fit method for Graph Neural Networks with graph structure

pub fn fit_graph( self, adjacency: &ArrayView2<'_, i32>, node_features: &ArrayView2<'_, Float>, node_labels: &Array2<i32>, ) -> SklResult<GraphNeuralNetwork<GraphNeuralNetworkTrained>>

Fit the GNN using graph structure, node features, and node labels

impl GraphNeuralNetwork<GraphNeuralNetworkTrained>

pub fn predict_graph( &self, adjacency: &ArrayView2<'_, i32>, node_features: &ArrayView2<'_, Float>, ) -> SklResult<Array2<i32>>

Predict node labels using the trained GNN

pub fn hidden_dim(&self) -> usize

Get the hidden dimension

pub fn num_layers(&self) -> usize

Get the number of layers

Trait Implementations

impl<S: Clone> Clone for GraphNeuralNetwork<S>

fn clone(&self) -> GraphNeuralNetwork<S>

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<S: Debug> Debug for GraphNeuralNetwork<S>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for GraphNeuralNetwork<Untrained>

fn default() -> Self

Returns the “default value” for a type.

impl Estimator for GraphNeuralNetwork<Untrained>

type Config = ()

Configuration type for the estimator

type Error = SklearsError

Error type for the estimator

type Float = f64

The numeric type used by this estimator

fn config(&self) -> &Self::Config

Get estimator configuration

fn validate_config(&self) -> Result<(), SklearsError>

Validate estimator configuration with detailed error context

fn check_compatibility( &self, n_samples: usize, n_features: usize, ) -> Result<(), SklearsError>

Check if estimator is compatible with given data dimensions

fn metadata(&self) -> EstimatorMetadata

Get estimator metadata