Predicting Parkinson’s Disease Progression: Evaluation of Ensemble Methods in Machine Learning

Table 4

Methods’ comparisons.

Method	Measure	MAE	RMSE	R^2	Computation time (ms)
NN	Motor-UPDRS	0.977	2.3836	0.7191	1072250
	Total-UPDRS	0.951	2.3135	0.7343	1043529
MLR	Motor-UPDRS	0.997	2.4142	0.6972	8953573
	Total-UPDRS	0.987	2.3911	0.7094	8845565
SVR	Motor-UPDRS	0.721	1.4942	0.8143	6743563
	Total-UPDRS	0.689	1.4526	0.8192	6633586
ANFIS	Motor-UPDRS	0.771	1.7047	0.7854	1534643
	Total-UPDRS	0.743	1.6062	0.7984	1525675
HSLSSVR	Motor-UPDRS	—	0.8158	—	—
	Total-UPDRS	—	0.8004	—	—
SOM + SVR	Motor-UPDRS	0.6340	0.5921	0.8518	538643
	Total-UPDRS	0.6421	0.6039	0.8421	535623
DBN	Motor-UPDRS	0.7645	1.6112	0.7914	974246
	Total-UPDRS	0.7321	1.5744	0.7996	964633
HGPA + SOM + SVR ensemble	Motor-UPDRS	0.5540	0.4116	0.9139	417435
	Total-UPDRS	0.5565	0.4179	0.9058	394352
HGPA + EM + SVR ensemble	Motor-UPDRS	0.5594	0.4165	0.9130	372223
	Total-UPDRS	0.5665	0.4186	0.9018	363422