Table 6 Classification and loss metrics for the best four selection methods (In %).

From: Explainable artificial intelligence approaches for COVID-19 prognosis prediction using clinical markers

Algorithm	Accuracy	F1-score	AUC	AP	MCC	Log loss	Jaccard score	Hamming loss
Mutual Information
Random forest	91	90	0.97	0.98	0.80	3.19	0.86	0.09
Logistic regression	87	87	0.93	90.96	0.73	4.47	0.80	0.12
Decision Tree	87	87	0.9	0.9	0.73	4.34	0.81	0.12
KNN	78	77	0.77	0.79	0.53	7.54	0.69	0.21
Adaboost	89	89	0.96	0.98	0.77	3.70	0.83	0.10
Catboost	91	91	0.97	0.98	0.81	3.07	0.86	0.08
Lightgbm	90	89	0.97	0.98	0.78	3.45	0.85	0.1
Xgboost	91	91	0.97	0.98	0.81	3.07	0.86	0.08
Stacking	90	89	0.96	0.98	0.78	3.58	0.84	0.10
Hard-Voting	95	95	0.98	0.96	0.89	1.66	0.93	0.04
Soft-Voting	94	94	0.98	0.99	0.87	1.918	0.92	0.88
Bat Algorithm
Random forest	90	89	0.95	0.97	0.77	3.45	0.86	0.1
Logistic regression	80	78	0.92	0.96	0.60	7.03	0.71	0.20
Decision Tree	87	85	0.92	0.95	0.71	4.47	0.82	0.12
KNN	77	76	0.79	0.85	0.53	7.00	0.695	0.22
Adaboost	88	87	0.95	0.97	0.73	4.09	0.83	0.11
Catboost	90	88	0.95	0.92	0.76	3.58	0.85	0.10
Lightgbm	92	91	0.95	0.97	0.81	2.81	0.88	0.08
Xgboost	92	91	0.95	0.98	0.81	2.68	0.89	0.07
Stacking	92	91	0.95	0.97	0.81	2.8	0.88	0.08
Hard-Voting	93	92	0.97	0.94	0.84	2.55	0.88	0.07
Soft-Voting	91	90	0.97	0.98	0.81	3.07	0.86	0.08
Flower Pollination Algorithm
Random forest	87	86	0.95	0.97	0.73	4.47	0.81	0.12
Logistic regression	83	83	0.92	0.95	0.67	5.88	0.74	0.17
Decision Tree	83	82	0.9	0.94	0.65	6.01	0.74	0.17
KNN	77	76	0.77	0.8	0.51	6.05	0.67	0.23
Adaboost	84	84	0.93	0.96	0.67	5.32	0.77	0.15
Catboost	88	87	0.95	0.97	0.75	4.22	0.81	0.12
Lightgbm	89	88	0.95	0.97	0.76	3.83	0.83	0.11
Xgboost	86	85	0.94	0.96	0.69	4.98	0.78	0.144
Stacking	87	87	0.94	0.97	0.75	4.34	0.81	0.12
Hard-Voting	85	84	0.93	0.87	0.68	5.11	0.79	0.14
Soft-Voting	86	84	0.93	0.96	0.69	4.98	0.80	0.14
Jaya Algorithm
Random forest	89	87	0.94	0.97	0.75	3.83	0.84	0.11
Logistic regression	80	78	0.92	0.96	0.60	7.05	0.71	0.20
Decision Tree	85	83	0.9	0.94	0.66	5.11	0.80	0.14
KNN	74	72	0.82	0.89	0.97	9.08	0.65	0.26
Adaboost	89	87	0.95	0.97	0.73	3.96	0.84	0.11
Catboost	89	87	0.94	0.97	0.74	3.96	0.84	0.11
Lightgbm	90	88	0.94	0.97	0.77	3.45	0.865	0.1
Xgboost	90	88	0.95	0.97	0.75	3.58	0.86	0.10
Stacking	89	88	0.94	0.97	0.75	3.70	0.855	0.10
Hard-Voting	89	88	0.95	0.89	0.76	3.96	0.82	0.1148
Soft-Voting	89	89	0.96	0.97	0.77	3.70	0.83	0.10

Back to article page

Table 6 Classification and loss metrics for the best four selection methods (In %).

Search

Quick links