E105-08 - Forschungsbereich Angewandte Statistik E101-03 - Forschungsbereich Scientific Computing and Modelling
-
Zeitschrift:
Computational Statistics & Data Analysis
-
ISSN:
0167-9473
-
Datum (veröffentlicht):
Apr-2022
-
Umfang:
20
-
Peer Reviewed:
Ja
-
Keywords:
Applied Mathematics; Computational Mathematics; Computational Theory and Mathematics; Prediction; Regression; Large sample size; Mean subspace; Nonparametric; Statistics and Probability
en
Abstract:
Neural networks are combined with sufficient dimension reduction methodology in order to remove the limitation of small p and n of the latter. NN-SDR applies when the dependence of the response Y on a set of predictors X is fully captured by the regression function , for an unknown function g and low rank parameter B matrix. It is shown that the proposed estimator is on par with competing sufficient dimension reduction methods, such as minimum average variance estimation and conditional variance estimation, in small p and n settings in simulations. Its main advantage is its scalability in regressions with large data, for which the other methods are infeasible.
en
Forschungsschwerpunkte:
Mathematical Methods in Economics: 65% Modelling and Simulation: 35%