Figure 5.

Prediction for fruit weight distribution tails. (A) All varieties except those in the large-fruit group (Fig. 2C), which is indicated by pink points, were used as the training set of the prediction model; the large-fruit group was used as the test set. (B) All varieties except those in the small-fruit group (Fig. 2C), which is indicated by blue points, were used as the training set; the small-fruit group was used as the test set. Grey points indicate the results of leave-one-out cross-validation. Predicted and observed values indicate log-transformed Weight. Pearson’s correlation coefficient (r) between predicted and observed values for the pink or blue points are shown on the plots. RR: linear ridge kernel regression, GAUSS: non-linear Gaussian kernel regression.