Optimized Design of Neural Networks for a River Water Level Prediction System

Miriam López Lineros; Antonio Madueño Luna; Pedro M Ferreira; Antonio E Ruano

doi:10.3390/s21196504

Optimized Design of Neural Networks for a River Water Level Prediction System

Sensors (Basel). 2021 Sep 29;21(19):6504. doi: 10.3390/s21196504.

Authors

Miriam López Lineros¹, Antonio Madueño Luna², Pedro M Ferreira³, Antonio E Ruano^{4

5}

Affiliations

¹ Design Engineering Department, University of Seville, 41013 Seville, Spain.
² Aerospace Engineering and Fluid Mechanical Department, University of Seville, 41013 Seville, Spain.
³ LASIGE, Departamento de Informática, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal.
⁴ Faculty of Science and Technology, University of Algarve, 8005-139 Faro, Portugal.
⁵ IDMEC, Instituto Superior Técnico, Universidade de Lisboa, 1049-001 Lisbon, Portugal.

Abstract

In this paper, a Multi-Objective Genetic Algorithm (MOGA) framework for the design of Artificial Neural Network (ANN) models is used to design 1-step-ahead prediction models of river water levels. The design procedure is a near-automatic method that, given the data at hand, can partition it into datasets and is able to determine a near-optimal model with the right topology and inputs, offering a good performance on unseen data, i.e., data not used for model design. An example using more than 11 years of water level data (593,178 samples) of the Carrión river collected at Villoldo gauge station shows that the MOGA framework can obtain low-complex models with excellent performance on unseen data, achieving an RMSE of 2.5 × 10^-3, which compares favorably with results obtained by alternative design.

Keywords: Multi-Objective Genetic Algorithm; artificial neural networks; river stage data.

MeSH terms

Neural Networks, Computer*
Rivers*
Water
Water Quality

Substances

Water