The deadwood model leverages boosted regression trees with inputs such as administrative linkage data, frame data, and historical response information as inputs, to produce a propensity score representing a relative likelihood of a farm operation being out of business. Common tree splits were identified using the model and combined with expert knowledge to develop a recurring process for deadwood clean up.
Date
Oct 2022
Source URL
Organization Type
Government