|
| Graph classification | Definition | Characteristic | Task |
|
| Isomorphic graph | , where V represents the node set and E represents the relationship set | ,, its node and relationship type are single, which can only represent binary relationship data | Link prediction, node classification, and clustering |
| Heterogeneous graph | , where V represents the node set and E represents the relationship set | , which has more than one node and relationship, and can only represent binary relationship data | Link prediction, node classification, and clustering |
| Heterogeneous hypergraph | , where V represents the node set and E represents the relationship set; for any superedge | , its node types are diverse, which can represent hyper-relational data, and the relationship usually has no semantics | Link prediction, node classification, and clustering |
| Knowledge hypergraph | , where V represents the entity set and E represents the hyper-relational set; the basic unit is the hyper-relational fact facta = (e, 1, 2,..., n) | ,, with various types of entities and relationships; it can represent hyper-relational data and rich relational semantics | Link prediction and hyper-relational fact judgment |
|
