Thumbnail Image
Publication

Feedforward neural networks from a statistical-modelling perspective

Date
2023-07-01
Abstract
Feedforward neural networks (FNNs) have many similarities to the models typically used in statistical modelling. Although they are often viewed as “black boxes”, when embedded in a statistical framework, these models can become more interpretable. A statistical-modelling approach to FNNs is demonstrated, with significance testing and covariate-effect plots used for explainability.
Supervisor
Description
Publisher
TU Dortmund University
Citation
Proceedings of the 37th International Workshop on Statistical Modelling (IWSM 2023), pp. 214-218
Collections
Mathematics & Statistics
Show all metadata
Files
Thumbnail Image
McInerney_2023_Feedforward
Adobe PDF, 953.32 KB
Keywords
neural networks, hypothesis testing, p-values, covariate effects
ULRR Identifiers
https://hdl.handle.net/10344/31678
 https://doi.org/10.34961/19313
Funding code
Funding Information
Sustainable Development Goals
External Link
License
Attribution-NonCommercial-ShareAlike 4.0 International
cbn
Embedded videos