Diabetes Bayesian Network with 413 Nodes and 602 Arcs

413 nodes and 602 arcs define a discrete Bayesian network for modeling diabetes. The network contains 429,409 parameters and was developed by S. Andreassen, R. Hovorka, J. Benn, K. G. Olesen, and E. R. Carson, with a related paper published in 1991.

Use Cases

• Probabilistic inference for diabetes-related variables based on the network's conditional dependencies.
• Developing clinical decision support tools for insulin adjustment based on the model's structure.
• Benchmarking structure learning algorithms against this established, large-scale network.
• Teaching Bayesian network concepts using a real-world, medically-grounded example.

Strengths

• Large-scale structure with 413 nodes and 602 arcs, providing a detailed probabilistic model.
• Well-defined provenance with clear authorship and a cited academic publication.
• Specific network statistics provided, including average Markov blanket size (3.97) and maximum in-degree (2).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and sample data are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

bnlearn Bayesian Network Repository, based on work by S. Andreassen et al.

Collection Method

Model-based approach for insulin adjustment, likely derived from clinical knowledge and data.

Time Range

null

Freshness

null

Geography

null