Unveiling the sound of the cognitive status: Machine Learning-based speech analysis in the Alzheimer’s disease spectrum

Table 10 Regression metrics obtained in the prediction of the composite cognitive scores

Composite	Model	RMAE (%)	MAE	Correlation	EV (%)
Attention	RF	18.0	0.474	0.468	21.8
	XGB	18.0	0.475	0.464	21.6
	GA-SVM	18.8	0.495	0.407	14.9
	GA-KNN	19.2	0.506	0.352	11.4
Executive function	RF	17.7	0.486	0.552	30.1
	XGB	17.4	0.477	0.563	31.7
	GA-SVM	18.3	0.502	0.517	26.1
	GA-KNN	19.2	0.527	0.450	19.9
Language	RF	18.1	0.466	0.548	29.9
	XGB	17.8	0.458	0.569	32.3
	GA-SVM	19.7	0.508	0.493	21.4
	GA-KNN	19.9	0.511	0.418	17.1
Memory	RF	18.5	0.534	0.494	24.3
	XGB	18.3	0.530	0.499	24.9
	GA-SVM	19.6	0.568	0.407	14.9
	GA-KNN	20.2	0.584	0.354	10.2
Visuospatial function	RF	18.1	0.521	0.565	31.7
	XGB	17.9	0.515	0.574	33.0
	GA-SVM	20.1	0.578	0.498	22.3
	GA-KNN	20.0	0.577	0.430	18.1

Abbreviations: MAE, mean absolute error; RMAE, relative MAE (described in the “Experimental setup” section); EV, explained variance; RF, random forest; XGB, extreme gradient boosting; GA-SVM, genetic algorithm- support vector machine; GA-KNN, genetic algorithm-K-nearest

ISSN: 1758-9193