Compare KNN and M5 Decision Tree Models in Anticipation of Evaporation and Comparison With Empirical Equations (Case Study of Birjand)

Evaporation from the soil surface is an important component of the water balance in the arid and semi-arid areas .In this study, to estimate the evaporation from open water surface, the empirical equations of Hefner, Shahtin, Marciano, Tichomirof, Ivanof, America Land Development Authority and Meyer have been used and modified in the study area. KNN model and M5 decision tree model were built ones data mining method to estimate the evaporation from open water surface. In so doing, data from meteorological stations in Birjand, Iran were used for finding the best effective input data on evaporation. Meanwhile, Gamma test was conducted to find the best combination of input parameters for estimating evaporation. To compare the results, the statistics of root mean square error (RMSE), correlation coefficient (R2) and mean absolute error (MAE) were analyzed. The results showed that M5 decision tree model reached the most optimum results than KNN model in most cases but the modified Meyer’s equation for Birjand area represented better estimation of open water surface evaporation due to 0.81 correlation coefficient and 2.06 errors.