Evaluation of LSSVM Model to Estimation of Evaporation and its Comparison with Empirical Models

In this research, we used the meteorological parameters of Birjand Plain in South Khorasan Province during the 16-year period in order to evaluate the gamma test and to compare the accuracy of the model of least squares of the machine and experimental models to estimate evaporation. Using the gamma test method, among the parameters affecting evaporation, the optimal input parameters were determined for modeling the evapotranspiration from among 90 specified components. Seven superior components were obtained from other combinations. Then, the optimal combination was then evaluated using the least squares model of backup vector with various kernel functions and empirical methods. The results showed that the performance of the LSSVM-poly (polynomial) model, in combination with minimum temperature, average temperature, wind speed and sunny hours with performance indices (R2 = 0.9995 and RMSE = 995) has a higher accuracy compared to the kernel functions and other experimental methods.