Reordering Features with Weights Fusion in Multiclass and Multiple-Kernel Speech Emotion Recognition

Table 1

Feature set.

Type	Feature	Statistic parameters
Prosodic feature	Pitch	Maximum (1), minimum (2), range (3), mean (4), Std (5), first quartile (6), median (7), third quartile (8), interquartile range (9)
	Energy	Maximum (10), minimum (11), range (12), mean (13), Std (14), first quartile (15), median (16), third quartile (17), interquartile range (18)
	Time	Total frames (19), voiced frames (20), unvoiced frames (21), ratio of voiced frames versus unvoiced frames (22), ratio of voiced frames versus total frames (23), ratio of unvoiced frames versus total frames (24)
Voice quality feature	Formant	F1:mean (25), Std (26), median (27)
		F2: mean (28), Std (29), median (30)
		F3: mean (31), Std (32), median (33)
Spectral feature	MFCC	12 MFCC (34–45)