GPSP: Graph Partition and Space Projection based Approach for Heterogeneous Network Embedding

TL;DR

GPSP introduces a novel method for embedding heterogeneous networks by partitioning into subnetworks and learning projective embeddings, leading to improved node classification and clustering performance.

Contribution

The paper presents a new approach combining graph partitioning and space projection for more effective heterogeneous network embedding.

Findings

GPSP outperforms state-of-the-art baselines in node classification.

GPSP achieves superior results in network clustering tasks.

Extensive experiments validate the effectiveness of GPSP.

Abstract

In this paper, we propose GPSP, a novel Graph Partition and Space Projection based approach, to learn the representation of a heterogeneous network that consists of multiple types of nodes and links. Concretely, we first partition the heterogeneous network into homogeneous and bipartite subnetworks. Then, the projective relations hidden in bipartite subnetworks are extracted by learning the projective embedding vectors. Finally, we concatenate the projective vectors from bipartite subnetworks with the ones learned from homogeneous subnetworks to form the final representation of the heterogeneous network. Extensive experiments are conducted on a real-life dataset. The results demonstrate that GPSP outperforms the state-of-the-art baselines in two key network mining tasks: node classification and clustering.