Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 28

Performance of hybrid systems on Pima Indian diabetics’ dataset.

Pima Indian diabetics
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	77.08	77.47	77.08	77.21	77.60	78.26	78.26	78.13
Sensitivity	72.78	72.51	72.53	72.93	73.63	74.63	74.63	74.51
Specificity	78.40	79.35	78.50	78.53	78.84	79.57	79.57	79.43
F-measure	75.86	76.74	75.90	76.01	76.45	77.45	77.45	77.30
Recall	77.08	77.47	77.08	77.21	77.60	78.26	78.26	78.13
Precision	76.51	76.97	76.49	76.65	77.09	77.85	77.85	77.71
	Basic MLP		Parameter optimized MLP
PAC	79.30	75.13	77.47	80.60	80.99	74.48	77.60	76.04
Sensitivity	70.71	65.34	69.41	76.59	73.14	64.06	70.87	65.79
Specificity	83.18	79.88	80.69	82.06	84.60	79.69	80.48	81.47
F-measure	79.09	74.93	76.97	79.97	80.83	74.34	77.16	76.02
Recall	79.30	75.13	77.47	80.60	80.99	74.48	77.60	76.04
Precision	78.99	74.81	76.90	80.22	80.75	74.24	77.13	76.00