Table 2 Performance comparison on IEMOCAP and MELD datasets.

From: MULTICAUSENET temporal attention for multimodal emotion cause pair extraction

Dataset

Method

Neutral

Happy

Sad

Angry

Excited

Frustrated

Surprise

Fear

Disgust

WF1

IEMOCAP

DialogueRNN18

59.91

32.83

76.20

64.21

71.83

60.94

-

-

-

55.43

DialogueGCN17

56.76

50.87

75.76

60.26

71.71

60.04

56.41

IterativeERC20

63.31

52.17

75.19

62.45

71.23

58.92

57.42

QMNN21

54.29

38.71

67.30

61.58

66.71

64.19

53.14

MMGCN19

63.73

41.34

77.67

67.00

73.33

61.32

59.11

MM-DFN22

65.42

43.22

79.98

71.77

73.56

68.33

62.56

MVN23

63.88

57.75

74.30

66.96

71.50

65.21

61.14

UniMSE24

EmoCaps2

65.48

70.91

84.06

67.99

78.41

64.76

70.15

GA2MIF25

70.38

45.15

83.50

71.29

74.99

65.49

67.46

MALN26

65.10

54.50

80.80

70.10

78.00

70.40

68.01

MultiEMO27

66.08

64.77

85.49

71.88

77.31

70.10

71.59

MultiCauseNet [Our]

70.51

74.51

83.21

71.61

77.52

71.95

73.02

MELD

DialogueRNN18

55.51

25.33

46.76

48.59

2.00

10.33

41.14

DialogueGCN17

52.95

26.32

42.03

46.37

1.98

12.37

40.67

IterativeERC20

55.95

21.62

49.88

52.65

5.31

21.24

41.78

QMNN21

53.18

15.50

42.17

51.76

1.00

39.33

MM-DFN22

54.24

53.78

47.82

35.95

MVN23

52.44

20.82

44.55

53.18

12.70

23.50

38.56

UniMSE24

43.52

58.54

62.19

1.03

-

40.90

EmoCaps2

56.90

43.52

56.54

62.19

3.03

15.65

42.87

GA2MIF25

52.10

28.18

49.52

48.08

39.89

MALN26

65.55

41.00

55.00

59.60

23.20

20.33

49.62

MultiEMO27

63.15

39.51

56.41

59.98

30.67

42.34

47.08

MultiCauseNet [Our]

66.98

51.29

59.41

63.21

31.32

35.43

53.67

  1. Significant values are in bold.