Module signed_directed 

Signed and directed graph learning with specialised embeddings.

This module provides:

• types: Core data structures (SignedGraph, DirectedGraph, config/result types).
• signed_spectral: Signed Laplacian, SPONGE embedding, ratio-cut clustering, and status-theory scores.
• directed_embedding: HOPE and APP directed graph embeddings.
• signed_gcn: Balance-theory Signed GCN (Derr et al. 2018).

Re-exports

pub use directed_embedding::app_embedding;

pub use directed_embedding::hope_embedding;

pub use directed_embedding::stationary_distribution;

pub use signed_gcn::predict_sign;

pub use signed_gcn::SgcnLayer;

pub use signed_gcn::SgcnModel;

pub use signed_spectral::negative_laplacian;

pub use signed_spectral::positive_laplacian;

pub use signed_spectral::signed_laplacian;

pub use signed_spectral::signed_ratio_cut;

pub use signed_spectral::sponge_embedding;

pub use signed_spectral::status_score;

pub use types::DirectedEdge;

pub use types::DirectedEmbedConfig;

pub use types::DirectedGraph;

pub use types::EmbeddingResult;

pub use types::SignedEdge;

pub use types::SignedEmbedConfig;

pub use types::SignedGraph;

Modules

directed_embedding

Directed graph embedding algorithms: HOPE (Ou et al. 2016) and APP (Zhou et al. 2017).

signed_gcn

Signed Graph Convolutional Network (SGCN) — Derr et al. 2018.

signed_spectral

Signed spectral embedding: Signed Laplacian, SPONGE (Cucuringu 2019), signed ratio-cut clustering, and status-theory score (Leskovec 2010).

types

Core types for signed and directed graph learning.