Ali Heidar Nasrolahi; yaser sabzevari
Abstract
In this research, in order to simulate the underground water level of Khorramabad plain, the performance of hybrid models of bat support vector regression, bat support vector regression, gray wolf support vector regression for four piezometric wells of Sarab Pardah, Naservand, Sally and Baba Hossein ...
Read More
In this research, in order to simulate the underground water level of Khorramabad plain, the performance of hybrid models of bat support vector regression, bat support vector regression, gray wolf support vector regression for four piezometric wells of Sarab Pardah, Naservand, Sally and Baba Hossein Bridge, which have homogeneous statistics. and it was done without missing data. For modeling, rainfall (P), temperature (T) and underground water level (H) and withdrawal from water resources (q) have been used in the monthly report of the models during the period of 1380-1400. It should be noted that for modeling, 80% of the data is chosen for training and the remaining 20% for testing, randomly, which covers a wide range of data types. Correlation coefficient (R), root mean square error (RMSE), mean absolute value of error (MAE) and Sutcliffe coefficient of vitality (NS) were used to evaluate and compare the performance of the models. The results showed that the combined structure provides better performance than other structures in all the investigated models. Also, the results showed that the wavelet support vector regression model based on the evaluation indicators, in the piezometric well of Sarab Pardah has R=0.978, RMSE=0.221 m, MAE=0.011 m, NS=0.985 and also in the piezometric well of Naservand has R=0.981 . 010 m, NS=0.986 and finally piezometric well, Baba Hossein Bridge has 5 R=0.98, RMSE=0.101 m, MAE=0.007 m, NS=0.995, compared to other models, it can be used to create a favorable result.
Geomorphology
hojatolah younesi; ahmad godarzi; Masoud Shakarami
Abstract
Today, hybrid models of artificial intelligence are considered as a suitable method for simulating hydrological phenomena, including quantitative estimation of river flow. For this purpose, there are various approaches in hydrology to estimate the flow rate of rivers, of which artificial intelligence ...
Read More
Today, hybrid models of artificial intelligence are considered as a suitable method for simulating hydrological phenomena, including quantitative estimation of river flow. For this purpose, there are various approaches in hydrology to estimate the flow rate of rivers, of which artificial intelligence models are the most important. Therefore, in this study, the performance of support vector-wavelet regression, backup vector-gray wolf regression and bat-support vector regression models to simulate the flow of Kashkan river located in Lorestan province during the statistical period of 2010-2011 in the daily time scale were analyzed. The criteria of correlation coefficient, root mean square error and mean absolute value of error and bias were selected for evaluation and performance of the models. The results showed that the hybrid models have acceptable results in simulating the river discharge. Comparison of models also showed that the support-wavelet vector regression model in the validation stage showed values of R2 = 0.960, RMSE = 0.045, MAE = 0.024, NS = 0.968 and BIAS = 0.001 in predicting daily river flow. . Overall, the results showed that the use of hybrid support-wavelet regression model can be useful in predicting daily discharge.
