Statistical and incremental methods for neural models selection

Authors:
Slim Abid;Mohamed Chtourou;Mohamed Djemel
Affiliations:
Control and Energy Management Lab CEM LAB, National School of Engineering of Sfax, University of Sfax, B.P. 1173, 3038 Sfax, Tunisia;Control and Energy Management Lab CEM LAB, National School of Engineering of Sfax, University of Sfax, B.P. 1173, 3038 Sfax, Tunisia;Control and Energy Management Lab CEM LAB, National School of Engineering of Sfax, University of Sfax, B.P. 1173, 3038 Sfax, Tunisia
Venue:
International Journal of Artificial Intelligence and Soft Computing
Year:
2014

Citing 13
Cited 0

Introduction to artificial neural systems

Introduction to artificial neural systems
Diffusion approximations for the constant learning rate backpropagation algorithm and resistance to local minima

Neural Computation
On cross validation for model selection

Neural Computation
Control of Electrical Drives

Control of Electrical Drives
Artificial Neural Networks: Concepts and Theory

Artificial Neural Networks: Concepts and Theory
Regularization Learning and Early Stopping in Linear Networks

IJCNN '00 Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks (IJCNN'00)-Volume 4 - Volume 4
Improving Bayesian Regularization of ANN via Pre-training with Early-Stopping

Neural Processing Letters
The general inefficiency of batch training for gradient descent learning

Neural Networks
New training strategies for constructive neural networks with application to regression problems

Neural Networks
Optimal convergence of on-line backpropagation

IEEE Transactions on Neural Networks
A general backpropagation algorithm for feedforward neural networks learning

IEEE Transactions on Neural Networks
Gradient methods for the optimization of dynamical systems containing neural networks

IEEE Transactions on Neural Networks
Training feedforward networks with the Marquardt algorithm

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

This work presents two methods of selection of neural models for identification of dynamic systems. Initially, a strategy of selection based on statistical tests, which relates to training and generalisation performances of a neural model is analysed. In the second time, a new constructive approach of neural model selection, which the training begins with minimal structure and then incrementally adds new hidden units and/or layers, is described. The simulation and the application of these methods for selection of neural models are also considered.