A wawelet based heuristic to dimension Neural Networks for simple signal approximation

Gabriele Colombini, Davide Sottara, Luca Luccarini, Paola Mello

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Before training a feed forward neural network, one needs to define its architecture. Even in simple feed-forward networks, the number of neurons of the hidden layer is a fundamental parameter, but it is not generally possible to compute its optimal value a priori. It is good practice to start from an initial number of neurons, then build, train and test several different networks with a similar hidden layer size, but this can be excessively expensive when the data to be learned are simple, while some real-time constraints have to be satisfied. This paper shows a heuristic method for dimensioning and initializing a network under such assumptions. The method has been tested on a project for waste water treatment monitoring.

Original languageEnglish (US)
Title of host publicationNeural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets
PublisherIOS Press
Pages110-115
Number of pages6
ISBN (Print)9781607500728
DOIs
StatePublished - Jan 1 2009
Externally publishedYes

Publication series

NameFrontiers in Artificial Intelligence and Applications
Volume204
ISSN (Print)0922-6389

Fingerprint

Neurons
Neural networks
Heuristic methods
Feedforward neural networks
Water treatment
Wastewater
Monitoring

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Colombini, G., Sottara, D., Luccarini, L., & Mello, P. (2009). A wawelet based heuristic to dimension Neural Networks for simple signal approximation. In Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets (pp. 110-115). (Frontiers in Artificial Intelligence and Applications; Vol. 204). IOS Press. https://doi.org/10.3233/978-1-60750-072-8-110

A wawelet based heuristic to dimension Neural Networks for simple signal approximation. / Colombini, Gabriele; Sottara, Davide; Luccarini, Luca; Mello, Paola.

Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets. IOS Press, 2009. p. 110-115 (Frontiers in Artificial Intelligence and Applications; Vol. 204).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Colombini, G, Sottara, D, Luccarini, L & Mello, P 2009, A wawelet based heuristic to dimension Neural Networks for simple signal approximation. in Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets. Frontiers in Artificial Intelligence and Applications, vol. 204, IOS Press, pp. 110-115. https://doi.org/10.3233/978-1-60750-072-8-110
Colombini G, Sottara D, Luccarini L, Mello P. A wawelet based heuristic to dimension Neural Networks for simple signal approximation. In Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets. IOS Press. 2009. p. 110-115. (Frontiers in Artificial Intelligence and Applications). https://doi.org/10.3233/978-1-60750-072-8-110
Colombini, Gabriele ; Sottara, Davide ; Luccarini, Luca ; Mello, Paola. / A wawelet based heuristic to dimension Neural Networks for simple signal approximation. Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets. IOS Press, 2009. pp. 110-115 (Frontiers in Artificial Intelligence and Applications).
@inproceedings{b9bc7b8c51d44fb1a6bfb6dad2735a03,
title = "A wawelet based heuristic to dimension Neural Networks for simple signal approximation",
abstract = "Before training a feed forward neural network, one needs to define its architecture. Even in simple feed-forward networks, the number of neurons of the hidden layer is a fundamental parameter, but it is not generally possible to compute its optimal value a priori. It is good practice to start from an initial number of neurons, then build, train and test several different networks with a similar hidden layer size, but this can be excessively expensive when the data to be learned are simple, while some real-time constraints have to be satisfied. This paper shows a heuristic method for dimensioning and initializing a network under such assumptions. The method has been tested on a project for waste water treatment monitoring.",
author = "Gabriele Colombini and Davide Sottara and Luca Luccarini and Paola Mello",
year = "2009",
month = "1",
day = "1",
doi = "10.3233/978-1-60750-072-8-110",
language = "English (US)",
isbn = "9781607500728",
series = "Frontiers in Artificial Intelligence and Applications",
publisher = "IOS Press",
pages = "110--115",
booktitle = "Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets",

}

TY - GEN

T1 - A wawelet based heuristic to dimension Neural Networks for simple signal approximation

AU - Colombini, Gabriele

AU - Sottara, Davide

AU - Luccarini, Luca

AU - Mello, Paola

PY - 2009/1/1

Y1 - 2009/1/1

N2 - Before training a feed forward neural network, one needs to define its architecture. Even in simple feed-forward networks, the number of neurons of the hidden layer is a fundamental parameter, but it is not generally possible to compute its optimal value a priori. It is good practice to start from an initial number of neurons, then build, train and test several different networks with a similar hidden layer size, but this can be excessively expensive when the data to be learned are simple, while some real-time constraints have to be satisfied. This paper shows a heuristic method for dimensioning and initializing a network under such assumptions. The method has been tested on a project for waste water treatment monitoring.

AB - Before training a feed forward neural network, one needs to define its architecture. Even in simple feed-forward networks, the number of neurons of the hidden layer is a fundamental parameter, but it is not generally possible to compute its optimal value a priori. It is good practice to start from an initial number of neurons, then build, train and test several different networks with a similar hidden layer size, but this can be excessively expensive when the data to be learned are simple, while some real-time constraints have to be satisfied. This paper shows a heuristic method for dimensioning and initializing a network under such assumptions. The method has been tested on a project for waste water treatment monitoring.

UR - http://www.scopus.com/inward/record.url?scp=74349084032&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=74349084032&partnerID=8YFLogxK

U2 - 10.3233/978-1-60750-072-8-110

DO - 10.3233/978-1-60750-072-8-110

M3 - Conference contribution

AN - SCOPUS:74349084032

SN - 9781607500728

T3 - Frontiers in Artificial Intelligence and Applications

SP - 110

EP - 115

BT - Neural Nets WIRN09 - Proceedings of the 19th Italian Workshop on Neural Nets

PB - IOS Press

ER -