Table 8 Comparing the performance of the proposed method in this study with the previous studies

(Li et al., 2019)[31]

SAP Data

LDA, SVM, NB, ANN

AUC = 91.2

(Liu et al., 2013)[10]

Personal Data, Fundus Images, Genome Data

SVM MKL

AUC = 86.6

(Li et al., 2018)[11]

Visual Field Repots

SVM, RF, K-NN, CNN

Acc = 87.6, Sen = 93.2, Spe = 82.6

(Noronha et al., 2019)[25]

Fundus Image

SVM, NB

Acc = 92.65, Sen = 100, Spe = 92.0

(Yo and Hong, 2015)[28]

Clinical Variables

MLR, ANN

Acc = 84.0, Sen = 78.3, Spe = 85.9

(Li et al., 2020)[29]

Fundus Image, Medical History Data

RNN(ResNet101)

Acc = 96.5, Sen = 99.8, Spe = 99.9

(Kim et al., 2017)[30]

RNFL Thickness, VF test Parameter, General Ophthalmic Examination

RF, DT, SVM, KNN

Acc = 98, Sen =98.3, Spe = 97.5

(Acharya et al., 2017)[26]

Fundus Images

DT, QDA, LDA, SVM, KNN, PNN

Acc = 95.8

(Mookiah et al., 2012)[27]

Fundus Images

SVM

Acc = 95.0, Sen =93.33, Spe = 96.67

(Chai et al., 2018)[32]

Fundus Images, Clinical Data

Multi Branch Neural Network

Acc = 99.24, Sen = 97.91, Spe = 93.59

(Pathan et al., 2021)[33]

Fundus Images

ANN, SVM, AdaBoost

Acc = 98.0, Sen = 100, Spe = 97.0

This study

Extensive ophthalmologic examination and clinical data

DT, RF, ET, KNN, SVM, Stacking Ensemble

Acc = 83.56, Sen = 82.21, Spe = 81.32