Search

In the present paper, we propose a new approach for monthly streamflow prediction based on the extreme learning machine (ELM) and the metaheuristic bat algorithm (Bat-ELM). The performance of the Bat-ELM was compared to that of ELM, support vector regression (SVR), Gaussian process regression (GPR), multilayer perceptron neural network (MLPNN), and generalized regression neural network (GRNN). The proposed models were applied using data from three hydrometric stations located in the Cheliff Basin, Algeria. The results showed that the Bat-ELM was more satisfactory than the standalone models. The Bat-ELM achieved the highest numerical performance with correlation coefficient and Nash-Sutcliffe efficiency ranging from 0.927 to 0.973 and from 0.846 to 0.944, respectively, much higher than the respective values obtained using the MLPNN, GRNN, SVR, GPR and ELM approaches. The obtained results demonstrate that the Bat-ELM is an interesting alternative algorithm for predicting high and extreme streamflow. [ABSTRACT FROM AUTHOR]

The objective of this paper is to classify the most adequate frequency laws to estimate the extreme floods in northern Algeria in the large Isser, Chellif and Cotiers Algerois Ouest drainage basins. The first step is to determine which laws could best represent the data used. The second step is to classify selected laws using various chosen criteria such as: BIC, AICc, and Schwartz criterion based on a priori and a posteriori probability concepts. The classification in this step is carried out using the unique information collected from observing each site, then working in the information collected from adjoining sites. Results show the prominence of the Gamma distribution. Key words: frequency analysis, northern Algeria, extreme flood values, regional analysis, Bayesian criterion, Gamma distribution, model selection criterion. Ce travail tente de dormer une classification des lois frequentielles les plus adequates a l'estimation des quantiles extremes de crues du nord de l'Algerie represente par les grands bassins versants de : Isser, Chellif et Cotiers-algerois Ouest. Pour ce faire, la premiere etape consiste en la determination du type de lois qui peut mieux representer les series de donnees utilisees. La seconde etape s'etend sur l'etablissement d'un classement des lois retenues; cela est rendu possible grace a l'application de differents criteres de choix a savoir : le BIC (critere bayesien), l'AICc (critere d'information d'akaike modifie) et le critere de Schwarz base sur les concepts de probabilite; a priori et a posteriori. Dans cette etape, un classement est etabli en utilisant l'unique information tiree des observations de chaque site independamment, puis d'exploiter l'information tiree des sites voisins. Les resultats obtenus montrent une preeminence de la loi Gamma. Mots-cles: analyse frequentielle, nord de l'Algerie, valeurs extremes de crues, analyse regionale, critere bayesien, loi Gamma, critere de selection de modeles. [Journal translation], 1. Introduction L'evaluation des valeurs extremes de crues revet un grand interet en hydrologie pour la resolution des problemes relatifs al'occupation des sols et au dimensionnement des ouvrages hydrauliques. Durant [...]