Graph dataset

As explained in Graph Neural Networks in Action

A dataset representing nodes, edges, their relationships, and associated node or edge features.

A graph dataset represents data through nodes, edges, their relationships, and node or edge features. Designing its structure explicitly helps test data integrity, avoid technical debt, and vary graph rules systematically in a GNN pipeline. section 8.2.2

Build the dataset around two complementary descriptions:

• Schema: specifies the graph’s structure and rules. section 8.2.2
• Instance model: applies the schema to a subset of the actual data. section 8.2.2
• Choose a simple homogeneous structure with one node and edge type, or extend it to heterogeneous graphs with multiple node and edge types, metadata, or ontologies. section 8.2.2

At left is the relationship nomenclature for ER diagrams. At right is an example of a conceptual schema using an ER diagram.

Use the ER-diagram nomenclature and conceptual schema to make entities, relationships, and constraints explicit before implementation. section 8.2.2

How Graph Neural Networks in Action covers it

How deep it goes

Defines schemas, instance models, and graph types, but stops short of implementation-specific storage or training procedures.

Examples

Uses ER diagrams, one-node-type schemas, multi-node-type schemas, and ontology relationships between writers and literary works.

What it assumes

Assumes basic familiarity with nodes, edges, and features while defining schemas, instances, and graph types explicitly.

What you can do after

Evaluate and design graph schemas that support integrity checks and systematic GNN pipeline experiments.

How it connects

Connects graph datasets with schemas, instance models, ER diagrams, homogeneous graphs, heterogeneous graphs, ontologies, and property graphs.

liveBookmanning.com

Read Graph Neural Networks in Action on liveBook Get the book at manning.com

Related concepts

• NetworkXused by · NetworkX is an open-source Python library for creating, manipulating, analyzing, and visualizing graphs and networks.
• graphkind of · A graph is a mathematical structure consisting of vertices connected by edges that represent relationships or connections.

This explainer was generated by Manning's knowledge graph from the text of Graph Neural Networks in Action. The section links point to the parts of the book it draws on.