Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization

Table 26

Performance of hybrid systems on Parkinson dataset.

Parkinson’s disease
	Resampling	Without resampling	Resampling			Without resampling
			PSO	GSA	FA	PSO	GSA	FA
	Basic SVM		Parameter optimized SVM
PAC	91.28	86.15	91.28	91.28	91.28	87.69	87.69	87.69
Sensitivity	90.00	87.04	90.00	90.00	90.00	86.39	86.39	86.39
Specificity	100.00	81.82	100.00	100.00	100.00	96.15	96.15	96.15
F-measure	90.41	85.21	90.41	90.41	90.41	86.29	86.29	86.29
Recall	91.28	86.15	91.28	91.28	91.28	87.69	87.69	87.69
Precision	92.15	85.75	92.15	92.15	92.15	88.79	88.79	88.79
	Basic MLP		Parameter optimized MLP
PAC	96.41	91.28	96.92	96.92	96.41	92.31	92.31	93.85
Sensitivity	97.40	94.52	97.42	97.42	97.40	95.83	96.48	96.55
Specificity	92.68	81.63	95.00	95.00	92.68	82.35	81.13	86.00
F-measure	96.39	91.31	96.90	96.90	96.39	92.38	92.43	93.89
Recall	96.41	91.28	96.92	96.92	96.41	92.31	92.31	93.85
Precision	96.39	91.35	96.90	96.90	96.39	92.52	92.70	93.95