Random forests are a widely used tool for measuring and predicting non-linear relations. Due to their non-linear nature, it is not possible to easily estimate the variance of the predicted values. Recently introduced Bootstrap-like procedures offer possibilities to estimate variances and therefore confidence intervals for Random Forests. Furthermore, it is also possible to estimate conditional quantiles with Random forests instead of estimating the conditional mean.

Literature:

BREIMAN, L. (2001). Random forests. Mach. Learn. 45 5–32.

MEINSHAUSEN, N. (2006). Quantile regression forests. J. Mach. Learn. Res. 7 983–999.

WAGER, S., HASTIE, T. and EFRON, B. (2014). Confidence intervals for random forests: The jackknife and the infinitesimal jackknife. J. Mach. Learn. Res. 15 1625–1651.

ATHEY, S., TIBSHIRANI, J. and WAGER, S. (2019). Generalized Random Forests. The Annals of Statistics 45 5-32.

Kontakt: Konstantin Görgen