Discussion of 'Comparative Study of Time Series Models, Support Vector Machines, and GMDH in Forecasting Long-Term Evapotranspiration Rates in Northern Iran' by Afshin Ashrafzadeh, Ozgur KiSi, Pouya Aghelpour, Seyed Mostafa Biazar, and Mohammadreza Askarizad Masouleh

摘要

In the current discussion, the discussers have provided some additional points on modeling process of the evapotranspiration rate forecasting to improve the quality of the proposed methods in the original paper. The main points can be summarized as follows: For the selection of the best model, in addition to accuracy criteria, the simplicity of the developed model should be verified. For development of machine learning models, the use of appropriate lag in evapotranspiration rate forecasting should be considered for selection of the best input combinations. In design of the group method of data handling structures, the number of tuned parameters should be kept less than the number of training samples.