Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 10

Performance of hybrid systems on Cleveland dataset.

Cleveland
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	85.14	82.83	86.79	86.13	86.13	83.49	82.83	83.49
Sensitivity	83.33	82.18	84.91	84.74	84.35	82.38	82.18	82.38
Specificity	87.80	83.72	89.51	88.09	88.70	85.03	83.72	85.03
F-measure	85.04	82.77	86.71	86.06	86.05	83.41	82.77	83.41
Recall	85.14	82.83	86.79	86.13	86.13	83.49	82.83	83.49
Precision	85.36	82.8	87.01	86.27	86.33	83.60	82.88	83.60
	Basic MLP		Parameter optimized MLP
PAC	93.72	79.20	94.05	94.05	94.05	85.14	84.15	85.80
Sensitivity	93.45	80.23	93.49	93.49	93.49	84.79	84.11	84.57
Specificity	94.07	77.94	94.77	94.77	94.77	85.60	84.21	87.5
F-measure	93.72	79.18	94.05	94.05	94.05	85.11	84.12	85.74
Recall	93.72	79.20	94.05	94.05	94.05	85.14	84.15	85.80
Precision	93.73	79.18	94.07	94.07	94.07	85.16	84.16	85.91