Win95pts: A 76-Node Bayesian Network for Probabilistic Inference

OpenML hosts the win95pts_6 dataset, a discrete Bayesian network with 76 nodes and 112 arcs. The network contains 574 parameters and is sourced from the bnlearn Bayesian Network Repository. Its author and last update date are unknown.

Use Cases

• Benchmarking structure learning algorithms based on the known network topology of 76 nodes and 112 arcs
• Evaluating probabilistic inference performance based on the network's 574 parameters
• Studying Markov blanket properties based on the reported average size of 5.92
• Analyzing graph connectivity patterns based on the reported average degree of 2.95 and maximum in-degree of 7

Strengths

• Well-defined network structure with 76 nodes and 112 arcs, providing a concrete test case
• Explicit parameter count of 574, indicating a defined probabilistic model
• Reported graph statistics, including an average Markov blanket size of 5.92 and maximum in-degree of 7

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment for certain tasks
• Last update date is unknown; freshness unverified

Provenance

Source

bnlearn Bayesian Network Repository

Collection Method

Likely a synthetic or expert-constructed model for benchmarking, as referenced from a discrete-large repository.

Time Range

unknown

Freshness

unknown

Geography

unknown