“…However, when UAS derived VIs have been used, especially in small plots and comparing varieties or treatments, other VIs have appeared as good predictors. For example, commonly selected UAS derived VIs that have been good predictors, aside from NDVI [ 22 , 31 ], include Green-Normalized Difference Vegetation Index (GNDVI) [ 2 , 11 , 22 , 31 , 32 ], Visible Atmospherically Resistant Index (VARI) [ 33 – 35 ], and Normalized Difference Red Edge (NDRE) [ 11 , 22 , 36 ]. Varied modeling procedures for incorporating and evaluating different VIs in yield prediction have been reported, including hierarchical linear regression [ 29 ], multiple linear regression [ 22 , 33 ], multivariate regression modeling with other phenological matrices [ 21 ], random forest [ 11 , 32 , 37 ], support vector machine [ 32 ], and boosted regression tree [ 30 ].…”