Depression-level assessment from multi-lingual conversational speech data using acoustic and text features

EURASIP Journal on Audio, Speech, and Music Processing

Table 7 Best classification results for the development set of the English task. Multi-lingual methods annotated with ml. There are 23 non-depressed and 12 depressed subjects in the database. Avec 2013 feature set used for all results except Gemaps tab. Number of selected features are shown in parenthesis for each case. Improvement with ml-MRMR using the Turkish database is statistically significant compared to the baseline MRMR algorithm

Method	Classes	Precision	Recall	F1-score
*Baseline*	Non-depressed	0.76	0.96	0.85
*MRMR(20)*	Depressed	0.83	0.42	0.56
	Average	0.79	0.69	0.71
*Eng+Tr N*_min=5	Non-depressed	0.79	1	0.88
*ml-MRMR(100)*	Depressed	1	0.50	0.66
	Average	0.89	0.75	0.77
*Eng+Ger N*_min=5	Non-depressed	0.76	0.96	0.85
*ml-MRMR(20)*	Depressed	0.83	0.42	0.56
	Average	0.79	0.69	0.71
*GEMAPS*	Non-depressed	0.70	0.82	0.76
	Depressed	0.50	0.33	0.40
	Average	0.60	0.58	0.58