Loading...
Statistical Information-Criteria-Based Neural Network Input and Hidden Node Selection
Date
2022-07
Abstract
Feedforward neural networks (FNNs) have many similarities to the models typically used in statistical modelling. The calculation of an associated likelihood function opens the door to information-criteria-based variable and architecture selection. A novel model selection method is proposed using the Bayesian information criterion for FNNs, wherein the optimal weights for one model are carried over to the next.
Supervisor
Description
Publisher
EUT Edizioni Università di Trieste
Citation
Proceedings of the 36th International Workshop on Statistical Modelling (IWSM 2022), pp. 238-241
Collections
Files
ULRR Identifiers
Funding code
Funding Information
Sustainable Development Goals
External Link
License
Attribution-NonCommercial-ShareAlike 4.0 International