English

Adaptive stratified Monte Carlo using decision trees

Computation 2025-01-10 v1

Abstract

It has been known for a long time that stratification is one possible strategy to obtain higher convergence rates for the Monte Carlo estimation of integrals over the hyper-cube [0,1]s[0, 1]^s of dimension ss. However, stratified estimators such as Haber's are not practical as ss grows, as they require O(ks)\mathcal{O}(k^s) evaluations for some k2k\geq 2. We propose an adaptive stratification strategy, where the strata are derived from a a decision tree applied to a preliminary sample. We show that this strategy leads to higher convergence rates, that is, the corresponding estimators converge at rate O(N1/2r)\mathcal{O}(N^{-1/2-r}) for some r>0r>0 for certain classes of functions. Empirically, we show through numerical experiments that the method may improve on standard Monte Carlo even when ss is large.

Keywords

Cite

@article{arxiv.2501.04842,
  title  = {Adaptive stratified Monte Carlo using decision trees},
  author = {Nicolas Chopin and Hejin Wang and Mathieu Gerber},
  journal= {arXiv preprint arXiv:2501.04842},
  year   = {2025}
}

Comments

20 pages, 6 figures

R2 v1 2026-06-28T21:00:32.117Z