Skip to main content

Table 6 Within-corpus results for binary classification on the USC IEMOCAP database

From: Articulation constrained learning with application to speech emotion recognition

  SVM ACO ACL Best target Best group BCUAR
Arousal
/AA/ 65.81 67.43 69.18* 72.60 (LPH, Z) 70.28 (LPH) 65.44
/AE/ 63.67 64.24 65.40 66.95 (LPH, X) 65.74 (LPH) 63.65
/IY/ 65.27 66.36 65.85 67.90 (LIP, Y) 67.70 (LIP) 66.22
/UW/ 62.49 63.90 64.44 67.60 (LPH, Z) 66.50 (LIP) 69.03
FULL 72.96 72.58 73.52 75.04 (CHN, X) 75.12 (CHN) 50.47
Valence
/AA/ 59.29 60.47 60.68 64.69 (LIP, Z) 63.01 (CHW) 51.21
/AE/ 57.20 57.61 59.06* 61.49 (LPW, X) 60.02 (CHW) 54.83
/IY/ 59.61 60.36 61.34 62.85 (CHN, Y) 61.85 (LPW) 56.88
/UW/ 61.20 63.17 63.62 64.42 (CHW, X) 63.95 (LIP) 56.25
FULL 60.34 61.74 62.16 63.37 (LPH, Z) 62.95 (LPW) 60.45
  1. The UAR is expressed in percentage. The columns of the table represent results from several models: support vector machine (SVM), acoustic only model (ACO), articulation constrained learning (ACL), best target using ACL, best group of targets using ACL, and the by-chance unweighted average recall (BCUAR). *Statistically significant improvement (p < 0.05) over compared methods