Skip to main content
. 2021 Nov 23;9:649221. doi: 10.3389/fbioe.2021.649221

FIGURE 4.

FIGURE 4

Relative importance of different features of prediction for the full model. These plots show the weights of the different features in the VA prediction task 1, 3, and 6 months after laser treatment in the full model. The results were stratified according to the follow-up period and the points input into the algorithms; this figure shows only the relative importance of the baseline data. The red bar indicates the average importance of the feature in the blending algorithm. (A) Feature weights in the 1-month VA prediction. (B) Feature weights in the 3-month VA prediction. (C) Feature weights in the 6-month VA prediction. VA, visual acuity; RNEL, retinal neuroepithelial layer; ChT, choroidal thickness; CMT, central macular thickness; DLS, double-layer sign; PDT, photodynamic therapy; SRF, subretinal fluid; ICGA, indocyanine green angiography; PED, pigment epithelial detachment.