Fig. 2

Performance comparison plot measured in Pearson’s correlation of our AGL-EAT-Score and other machine learning-based models on the a CASF-2013, and b CASF-2016 benchmark datasets. Our model AGL-EAT-Score (highlighted in red color) scores. In CASF-2013 benchmark, our model achieved \(R_p = 0.845\) and the results of other methods obtained from prior research [59, 60, 66]. In CASF-2016 benchmark, our model achieved \(R_p = 0.873\) and the results of other methods obtained from prior research [57, 59, 67]