Forecasting of Lake Level by Soft Computing Approaches
[ X ]
Tarih
2024
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Springer Science and Business Media Deutschland GmbH
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
To ensure the sustainability and management of water resources, regularly monitoring the water levels in lakes, rivers, basins, dam reservoirs, etc. is a very important engineering task. Our freshwater resources are gradually decreasing due to the destruction of freshwater resources and climate change. For this reason, monitoring, modelling, and researching of freshwater resources, especially lakes, are increasingly important issue for nowadays. In this chapter, soft computing approaches are used to forecast of lake water levels at Beyşehir Lake, located in the central part of Turkey. To do this, three artificial neural network algorithms (Multilayer, Radial Basis and Generalized Regression), two heuristic algorithms (Model 5-Tree and Multivariate Adaptive Regression Spline), and a Support Vector Machines containing three different functions (Radial, Polynomial, and Linear) are used. In addition to being models used successfully in hydrological modelling of civil engineering, the changes in modelling performance with the number of iterations, kernel functions, optimization algorithms, and data input sets that constitute the internal dynamite of the techniques are investigated. The attained results show that through these multiple parameters, radial basis artificial neural networks are the most successful when compared with mean absolute error, root mean square error, coefficient of determination, Taylor diagrams and Violin plots. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
Açıklama
Anahtar Kelimeler
Artificial neural network, Heuristic algorithms, Lake-level forecasting, Support vector machines
Kaynak
Studies in Systems, Decision and Control
WoS Q Değeri
Scopus Q Değeri
Q2
Cilt
547
