Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 24

Performance of hybrid systems on thyroid dataset.

Thyroid
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	89.77	89.77	91.16	91.16	91.16	89.77	89.77	89.77
Sensitivity	90.70	93.88	97.50	97.50	97.50	93.88	93.88	93.88
Specificity	89.53	88.55	89.71	89.71	89.71	88.55	88.55	88.55
F-measure	89.27	89.31	90.61	90.61	90.61	89.31	89.31	89.31
Recall	89.77	89.77	91.16	91.16	91.16	89.77	89.77	89.77
Precision	89.84	90.16	91.78	91.78	91.78	90.16	90.16	90.16
	Basic MLP		Parameter optimized MLP
PAC	98.14	96.28	98.60	97.67	98.6	97.67	96.74	97.67
Sensitivity	98.18	93.85	98.21	94.83	98.21	95.45	93.94	95.45
Specificity	98.13	97.33	98.74	98.73	98.74	98.66	97.99	98.66
F-measure	98.13	96.28	98.60	97.68	98.6	97.68	96.75	97.68
Recall	98.14	96.28	98.60	97.67	98.6	97.67	96.74	97.67
Precision	98.14	96.28	98.60	97.69	98.6	97.69	96.76	97.69