Fig. 4.

Optimal model results. The abscissa reflects the optimal combination of parameters for models with different numbers of variables under different fitting methods, where 1 V-NN represents the use of a single variable (ICL size). 2 V-NN: ICL size, ATA; 3 V-NN: ICL size, ACD, ATA; 4 V-NN: ICL size, ACD, ATA, and LT; 11 V-NN/11 V-Linear: age, sex, spherical equivalent, expected spherical equivalent, WTW, ACD, ATA, STS, LT, ICL size, ICL optical power. NN represents the neural network fitting model. R² values are presented on the y-axis. WTW, white-to-white; ACD, anterior chamber depth; ATA, angle-to-angle; STS, sulcus-to-sulcus; LT, lens thickness; ICL, implantable collamer lens