Diabetes Bayesian Network with 413 Nodes for Insulin Adjustment Modeling

A Bayesian network model for diabetes management, containing 413 nodes and 602 arcs. The model was developed by S. Andreassen, R. Hovorka, J. Benn, K. G. Olesen, and E. R. Carson and presented at the 3rd Conference on Artificial Intelligence in Medicine in 1991. It has 429,409 parameters and is designed for a model-based approach to insulin adjustment.

Use Cases

• Probabilistic inference for insulin dosage adjustment based on the network's structure and parameters.
• Benchmarking Bayesian network learning algorithms using the published model topology.
• Studying causal relationships in diabetes pathophysiology using the defined arcs and nodes.
• Developing educational tools for clinical decision-making in endocrinology.

Strengths

• The model is a well-defined Bayesian network with 413 nodes and 602 arcs, providing a detailed structure.
• It has a substantial number of parameters (429,409) for probabilistic modeling.
• The model is explicitly cited in a peer-reviewed conference proceeding from 1991, establishing academic provenance.

Limitations

• Column-level documentation and sample data are unavailable, making field semantics and data format uncertain.
• Row count and file size are unknown, which limits suitability assessment for direct data analysis.
• The last update date is unknown; the model's alignment with contemporary medical knowledge is unverified.

Provenance

Source

S. Andreassen, R. Hovorka, J. Benn, K. G. Olesen, and E. R. Carson.

Collection Method

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

Time Range

null

Freshness

null

Geography

null