Hailfinder: A Bayesian Network for Severe Weather Forecasting
by B. Abramson, J. Brown, W. Edwards, A. Murphy, and R. L. Winkler.
arff
Available on 1 platform
Sign in to view source links and access this dataset
Description
56 nodes and 66 arcs define the Hailfinder Bayesian network, a probabilistic model for forecasting severe weather. The network contains 2656 parameters and was authored by B. Abramson, J. Brown, W. Edwards, A. Murphy, and R. L. Winkler, with research published in the International Journal of Forecasting in 1996.
Use Cases
Benchmarking Bayesian network structure learning algorithms based on the known network topology
Testing probabilistic inference algorithms based on the 2656 parameters and conditional probability tables
Studying decision support systems for meteorology based on the weather forecasting application
Educational demonstrations of Bayesian networks based on the documented structure and domain
Strengths
Well-documented network structure with 56 nodes and 66 arcs
Published academic provenance with a clear citation to peer-reviewed research
Specific network statistics provided, including an average Markov blanket size of 3.54 and maximum in-degree of 4
Limitations
Column-level documentation is absent; field semantics must be inferred after download
Row count is unknown, which may limit suitability assessment for some learning tasks