“…The p value of lack of fit was 0.7268, which also suggested a fitted model [ 28 ]. Besides, statistical indicators, including R 2 , Adj R 2 , Pred R 2 , and CV values, were 0.9752, 0.9504, 0.8935, and 3.18, respectively, and the difference of Pred R 2 and Adj R 2 was less than 0.2 (0.0569), indicating a good correlation between actual and predicted values [ 29 ]. The parameters, including A , B , C , D , AD , BC , BD , A 2 , B 2 , C 2 , and D 2 , exhibited remarkable differences ( p < 0.05) to SMGP yields, which was a similar tendency observed in a previously reported paper [ 30 ].…”