Fuzzy neural networks for water level and discharge forecasting with uncertainty

This paper proposes a new procedure for water level (or discharge) forecasting under uncertainty using artificial neural networks: uncertainty is expressed in the form of a fuzzy number. For this purpose, the parameters of the neural network, namely, the weights and biases, are represented by fuzzy...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Environmental modelling & software : with environment data news Jg. 26; H. 4; S. 523 - 537
Hauptverfasser: Alvisi, Stefano, Franchini, Marco
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 01.04.2011
Schlagworte:
Artificial neural networks
Bands
computer software
Discharge forecasting
environmental models
Fuzzy
Fuzzy logic
Fuzzy numbers
Fuzzy set theory
Fuzzy systems
Mathematical models
Neural networks
Uncertainty
Water level forecasting
Uncertainty
Water level forecasting
Fuzzy numbers
Artificial neural networks
Discharge forecasting
ISSN:1364-8152
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper proposes a new procedure for water level (or discharge) forecasting under uncertainty using artificial neural networks: uncertainty is expressed in the form of a fuzzy number. For this purpose, the parameters of the neural network, namely, the weights and biases, are represented by fuzzy numbers rather than crisp numbers. Through the application of the extension principle (Zadeh, 1965), the fuzzy number representative of the output variable is then calculated at each time step on the basis of a set of crisp inputs and fuzzy parameters of the neural network. The fuzzy parameters of the neural network are estimated from the modelling process, that is through a calibration procedure that imposes a constraint whereby for an assigned h-level the envelope of the corresponding intervals representing the outputs (forecasted levels or discharges, calculated at different points in time) must include a preset percentage of observed values. The application of the model to two specific cases and a comparison of the results with those provided by other data-driven models – Bayesian neural networks (Neal, 1992) and the Local Uncertainty Estimation Model (Shrestha and Solomatine, 2006) – show its effectiveness in estimating water levels or discharges under uncertainty. The fuzzy neural network enables us to define bands that are expected to contain given percentages of forecasted level/discharge values for each lead time selected; an analysis of the results obtained reveals that these bands (e.g., 99%, 95% and 90% uncertainty bands) generally have a slightly smaller width compared to the bands obtained using other data-driven models.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:1364-8152
DOI:10.1016/j.envsoft.2010.10.016