Lost In The Forest: Encoding Categorical Variables and the Absent Levels Problem
Helen L. Smith, Patrick J. Biggs, Nigel Peter French, Adam N. H. Smith, Jonathan C. Marshall
The paper addresses the absent levels problem in random forest models by proposing a novel ordinal encoding method using principal coordinates analysis, showing it is unbiased and performs as well as target-based approaches.
Data Mining and Knowledge Discovery · Apr 10, 2024Peer-reviewedBioinformatics