Skip to main content

Table 7 Best results after parameter tuning for arousal and valence tasks with according settings (feature set, feature selection method, and percentage of features). Top, overall best settings; Mid, overall best with only the EmoFt set; Bottom, overall best with only the EmoFt set and linear kernel SVM

From: Emotion in the singing voice—a deeperlook at acoustic features in the light ofautomatic classification

Dimension

UAR [%]

Setting

Overall best results

  

Arousal

61.7

ComParE, CC-FS (50 %), linear, C=0.05

Valence

52.9

ComParE, CC-FS (10 %), linear, C=1.0

EmoFt set best results

  

Arousal

60.1

EmoFt, CC-FS (50 %), RBF kernel, C=1.0

Valence

48.0

EmoFt, CC-FS (50 %), RBF kernel, C=1.0

EmoFt set, only linear kernel SVM, best results

  

Arousal

58.0

ComParE, SPKSTD-CC (30 %), linear, C=0.05

Valence

45.3

ComParE, all features, linear, C=1.0