Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 30

Performance of hybrid systems on BUPA liver disease dataset.

BUPA liver disease
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	63.48	70.14	66.09	66.67	63.48	70.72	70.14	70.14
Sensitivity	61.44	68.10	64.86	65.97	61.44	72.00	68.10	68.10
Specificity	65.10	71.18	67.01	67.16	65.10	70.20	71.18	71.18
F-measure	63.40	69.50	65.95	66.47	63.40	69.52	69.50	69.50
Recall	63.48	70.14	66.09	66.67	63.48	70.72	70.14	70.14
Precision	63.39	69.89	66.01	66.61	63.39	70.96	69.89	69.89
	Basic MLP		Parameter optimized MLP
PAC	68.12	71.59	65.80	73.33	70.43	72.75	73.04	71.59
Sensitivity	63.93	69.75	61.88	68.85	67.46	70.73	70.00	68.50
Specificity	72.84	72.57	70.12	78.40	73.30	73.87	74.88	73.39
F-measure	68.12	71.06	65.81	73.34	70.46	72.35	72.80	71.27
Recall	68.12	71.59	65.80	73.33	70.43	72.75	73.04	71.59
Precision	68.68	71.38	66.27	73.94	70.57	72.55	72.83	71.34