Enhancement of Detection of Diabetic Retinopathy Using Harris Hawks Optimization with Deep Learning Model

Table 2

Summary of the experimental results.

Metric ⟶method ↓	DNN	DNN-PCA	DNN-PCA-HHO	DT	DT-PCA	DT-PCA-HHO	KNN	KNN-PCA	KNN-PCA-HHO	NB	NB-PCA	NB-PCA-HHO	SVM	SVM-PCA	SVM-PCA-HHO	XGBoost	XGBoost-PCA	XGBoost-PCA-HHO
Accuracy	96.7	91.2	97	89.1	87	91.2	59.4	66	73.9	58	65	71.5	75	66	77.5	84.5	81	88
Precision	95.6	90.7	97	89.6	87	92.3	59.2	66	74.1	58	67	72.7	76	79	81.7	85.2	81	90
Recall	95.7	91.4	97	90.2	88	93.0	59.3	67	79.1	60	65	71.6	75	66	78.6	84.3	81	90
Sensitivity	91.1	88.3	91	86.9	85	90.5	60.9	64	72.1	61	63	72.5	84	86	88.2	76.5	82	84
Specificity	95.1	92.2	96	91.1	91	93.4	57.9	64	68.2	58	65	73.2	85	60	90.3	64.6	80	82