Performance of machine learning algorithms for dementia assessment: impacts of language tasks, recording media, and modalities

Table 3 List of acoustic features that are considered in this research

Type	Name	Functional	# of Features
Spectral features	MFCCs 0–14	Mean, kurt, skew, std	60
	ΔMFCCs 0–14	Mean, kurt, skew, std	60
	log Mel freq 0–7	Mean, kurt, skew, std	32
	Δlog Mel freq 0–7	Mean, kurt, skew, std	32
	LSP freq 0–7	Mean, kurt, skew, std	32
	Δ LSP freq 0–7	Mean, kurt, skew, std	32
Phonation and voice	F0	Mean, kurt, skew, std	4
Quality features	ΔF0	Mean, kurt, skew, std	4
	Jitter local	Mean, kurt, skew, std	4
	Δjitter local	Mean, kurt, skew, std	4
	Jitter DDP	Mean, kurt, skew, std	4
	Δjitter DDP	Mean, kurt, skew, std	4
	Shimmer	Mean, kurt, skew, std	4
	Δshimmer	Mean, kurt, skew, std	4
	Loudness	mean, kurt, skew, std	4
	Δloudness	Mean, kurt, skew, std	4
Speech features	Voicing prob.	Mean, kurt, skew, std	4
	Δvoicing prob.	Mean, kurt, skew, std	4
Total			296

ISSN: 1472-6947