Interpretable Approaches to Predict Evapotranspiration

In this short paper, we compare for the first time the performance of some approaches for evapotranspiration prediction, in terms of interpretability, accuracy and training time. The considered techniques are Decision Trees (DTs) and the Adaptive Network-based Fuzzy Inference System with fractional Tikhonov regularization (ANFIS-T), which are known as interpretable. Although interpretable to some extent, Support Vector Regression (SVR) was also included in the comparative analysis, since commonly used in precision agriculture. Experiments were performed on two publicly available datasets. ANFIS-T showed the highest level of interpretability, with comparable accuracy and reduced training time with respect to DTs.