A0944
Title: Asymptotic property for generalized random forests
Authors: Hiroshi Shiraishi - Keio University (Japan) [presenting]
Tomoshige Nakamura - Juntendo University (Japan)
Ryuta Suzuki - Keio University (Japan)
Abstract: The asymptotic property of the generalized random forests (GRF) estimator proposed in a prior study is discussed. It derives the asymptotic normality of the GRF estimator, but it does not explicitly derive the rate of convergence and asymptotic variance. The aim is to derive the rate of convergence and explicit form of the asymptotic variance of the GRF estimator. To overcome this problem, the objective function is approximated as a class of Nadraya-Watoson type statistics and its asymptotic normality is derived by some modification of other studies.
