Machine Learning for the Preliminary Diagnosis of Dementia

Table 5

Performance of the diagnostic models in the classification of normal, MCI, VMD, and dementia.

Algorithm	Class	Precision	Sensitivity	Specificity	F-measure
Random Forest	Normal	0.56	0.88	0.93	0.69
	MCI	0.70	0.57	0.93	0.62
	VMD	0.68	0.54	0.94	0.60
	Dementia	0.91	0.95	0.90	0.93
AdaBoost	Normal	0.55	0.84	0.93	0.67
	MCI	0.74	0.54	0.95	0.63
	VMD	0.63	0.55	0.93	0.59
	Dementia	0.89	0.94	0.88	0.92
LogitBoost	Normal	0.56	0.56	0.98	0.67
	MCI	0.73	0.73	0.89	0.65
	VMD	0.77	0.77	0.87	0.47
	Dementia	0.83	0.83	0.98	0.90
MLP	Normal	0.77	0.84	0.74	0.80
	MCI	0.65	0.74	0.89	0.69
	VMD	0.57	0.37	0.94	0.45
	Dementia	0.88	0.93	0.87	0.90
Naïve Bayes	Normal	0.56	0.84	0.94	0.67
	MCI	0.75	0.62	0.93	0.68
	VMD	0.70	0.72	0.93	0.71
	Dementia	0.95	0.92	0.95	0.93
SVM	Normal	0	0	1	0
	MCI	0.60	0.96	0.83	0.74
	VMD	0.85	0.56	0.98	0.67
	Dementia	0.91	0.97	0.90	0.94