Just add noise: Debiasing tree-based variable importance in mixed data
This story is from 2026-09-15. It is preserved in the archive; the latest stories are on the live feed.
arXiv:2609.14083v1 Announce Type: new Abstract: Variable importance scores from tree-based methods such as random forests favor continuous predictors over categorical ones. We present a theoretical analysis of this bias and propose a simple remedy: add a small amount of noise to each categorical pr…
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-09-15 04:00 · arXiv stat.ML
Just add noise: Debiasing tree-based variable importance in mixed data