Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 14

Performance of hybrid systems on Spect dataset.

Spect
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	88.39	81.65	88.39	88.39	88.39	82.77	82.40	82.77
Sensitivity	91.07	88.26	91.07	91.44	91.07	87.05	88.02	88.43
Specificity	74.42	55.56	74.42	73.33	74.42	60.47	58.00	58.82
F-measure	87.96	81.59	87.96	88.06	87.96	81.95	82.08	82.53
Recall	88.39	81.65	88.39	88.39	88.39	82.77	82.40	82.77
Precision	87.83	81.53	87.83	87.91	87.83	81.58	81.83	82.33
	Basic MLP		Parameter optimized MLP
PAC	88.39	79.40	89.51	89.51	89.14	82.40	83.52	85.02
Sensitivity	91.44	87.56	91.93	91.93	91.89	87.33	87.17	86.44
Specificity	73.33	50.00	77.27	77.27	75.56	58.70	63.41	74.19
F-measure	88.06	79.60	89.17	89.17	88.83	81.80	82.58	83.33
Recall	88.39	79.40	89.51	89.51	89.14	82.40	83.52	85.02
Precision	87.91	79.82	89.07	89.07	88.71	81.43	82.28	83.92