“…Feature selection is one of the effective ways to reduce dimensionality, which helps to reduce the risk of overfitting, improve the generalization ability of the model and save computational effort because only a fewer features are calculated ( Shilaskar and Ghatol, 2013 ). On the other hand, machine learning methods, such as support vector machines (SVM), logistic regression (LR), random forests (RF), K-nearest neighbors (KNN), and decision trees (DT), are widely used for classification and prediction of ophthalmic diseases, such as myopia and keratitis ( Tang et al, 2020 ; Herber et al, 2021 ), glaucoma, uveitis, cataract, and age-related macular degeneration ( Lin et al, 2020 ; Standardization of Uveitis Nomenclature SUN Working Group, 2021b ; Ting et al, 2021 ), and recently also for VKH classification ( Standardization of Uveitis Nomenclature SUN Working Group, 2021a ; Chang et al, 2021 ), because of their good classification performance in small datasets. These classifiers are often trained with hyperparameters, which need to be optimized to obtain the best classification performance.…”