“…The ability of artificial intelligence (AI) to portray the non‐linear relationship between inputs and ET o without depicting the complex process behind it led to various machine learning (ML) based ET o modelling research, especially those that utilized minimum, easily available, meteorological parameters. Most of this research was executed in regions where the essential meteorological data is unavailable and explored traditional ML to deep learning (DL) architectures such as artificial neural networks (ANN) and its derivatives (Ferreira & da Cunha, 2020; Kaya et al, 2021; Nagappan et al, 2020; Sowmya et al, 2020), support vector machine (SVM) and its hybrid variants (Chia, Huang, & Koo, 2020; Tikhamarine et al, 2020), adaptive neuro fuzzy inference systems (Gonzalez del Cerro et al, 2021; Üneş et al, 2020), gene expression programming (Kazemi et al, 2021), tree based soft computing methods (Fan et al, 2018; Rashid Niaghi et al, 2021; Salam & Islam, 2020), hybrid ML models (Chen et al, 2020; Kisi et al, 2021; Muhammad et al, 2021; Zhu et al, 2020), and ensemble of ML models (Wu et al, 2021). These investigations demonstrated excellence as well as superiority over traditional empirical approaches, when temperature and radiation data were included in the input set.…”