MF2Vec is a model proposed to overcome the limitations of existing heterogeneous graph neural networks (HGNNs), which rely on domain-specific, predefined metapaths. While existing methods focus only on simple aspects such as node types, MF2Vec extracts fine-grained paths through random walks, thereby ignoring predefined schemas and learning diverse aspects of nodes and relationships. The resulting multi-faceted vectors form homogeneous networks and generate node embeddings, which are then utilized for various tasks such as classification, link prediction, and clustering. Experimental results demonstrate that MF2Vec outperforms existing methods and provides a more flexible and comprehensive framework for complex network analysis.
