Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 16

Performance of hybrid systems on Spectf dataset.

Spectf
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	87.64	79.40	89.14	89.14	88.01	80.90	79.40	80.15
Sensitivity	92.92	87.56	94.71	94.71	92.96	87.10	87.56	88.04
Specificity	67.27	50.00	69.49	69.49	68.52	54.00	50.00	51.72
F-measure	87.77	79.60	89.39	89.39	88.10	80.56	79.60	80.34
Recall	87.64	79.40	89.14	89.14	88.01	80.90	79.40	80.15
Precision	87.93	79.82	89.80	89.80	88.20	80.28	79.82	80.56
	Basic MLP		Parameter optimized MLP
PAC	90.26	76.03	90.64	89.89	90.26	80.52	80.90	82.40
Sensitivity	93.15	85.24	93.58	93.12	93.55	87.74	84.85	82.35
Specificity	77.08	42.11	77.55	75.51	76.00	52.73	55.56	83.33
F-measure	90.11	76.19	90.53	89.77	90.19	80.52	79.31	77.56
Recall	90.26	76.03	90.64	89.89	90.26	80.52	80.90	82.40
Precision	90.02	76.35	90.46	89.69	90.13	80.52	78.81	82.55