Estimation of a regression function on a manifold by fully connected deep neural networks

Michael Kohler, Sophie Langer*, Ulrich Reif

*Corresponding author for this work

Research output: Working paper

3 Downloads (Pure)

Abstract

Estimation of a regression function from independent and identically distributed data is considered. The L2 error with integration with respect to the distribution of the predictor variable is used as the error criterion. The rate of convergence of least squares estimates based on fully connected spaces of deep neural networks with ReLU activation function is analyzed for smooth regression functions. It is shown that in case that the distribution of the predictor variable is concentrated on a manifold, these estimates achieve a rate of convergence which depends on the dimension of the manifold and not on the number of components of the predictor variable.
Original languageEnglish
PublisherarXiv.org
Number of pages39
Publication statusPublished - 2021
Externally publishedYes

Fingerprint

Dive into the research topics of 'Estimation of a regression function on a manifold by fully connected deep neural networks'. Together they form a unique fingerprint.

Cite this