Skip to main content
. 2021 Mar 1;12:506681. doi: 10.3389/fpls.2021.506681

Figure 11.

Figure 11

Qualitative results on non-redundant dataset using 188D. (A,D) Refer to the accumulated results of the 50-fold scatter plots, each of which corresponds to one time of random sample selection for training. (B,E) Gaussian distributions of the accumulated variable importance corresponding to twice random sample splits, respectively. (C,F) ACC line charts of different dimensions derived from feature 188D, with important variables incrementally added one by one according to the accumulated scores in a descending order.