Accuracy enhancement for monthly evaporation predicting model utilizing evolutionary machine learning methods

摘要

Evaporation is an important parameter for water resource management. In this article, two case studies with different climates were considered in the prediction of monthly evaporation. The optimization algorithms, namely shark algorithm (SA) and firefly algorithms (FFAs), were used to train the adaptive neuro-fuzzy interface system (ANFIS), multilayer perceptron (MLP) model and radial basis function (RBF) model for the prediction of monthly evaporation. The monthly weather data from two stations, Mianeh station and Yazd station, operated by the Iran Meteorological Service were used to examine the proposed models. In the quantitative analysis, the hybrid ANFIS-SA improved the MAE index over the ANFIS, RBF, MLP, RBF-SA, MLP-SA, RBF-FFA, MLP-FFA and ANFIS-SA up to 47% during training and to 51% during testing while examining Yazd station. It should be mentioned that the higher RSR and MAE were attained by the hybrid soft computing (ANN-FFA, RBF-FFA and ANFIS-FFA) models in two stations. The results proved that the developed ANFIS models that have been integrated with shark algorithms could be considered as a powerful tool for predicting evaporation.