Table 4 5-fold cross-validated paired t-tests comparing macro-average precision, recall, F1 scores, and accuracy of FinBERT-based models against the xFiTRNN model.

From: A hybrid self attentive linearized phrase structured transformer based RNN for financial sentence analysis with sentence level explainability

Embedding

Model

t-value

(A)

p-value

(A)

t-value

(P)

p-value

(P)

t-value

(R)

p-value

(R)

t-value

(F1)

p-value

(F1)

FinBERT

1 Hidden Layer NN

42.57

3.24e–11

30.14

5.53e–10

38.79

2.13e–11

35.23

9.83e–11

2 Hidden Layers NN

47.83

1.52e–12

34.92

2.34e–11

45.31

3.41e–12

40.15

7.63e–12

3 Hidden Layers NN

54.27

4.83e–14

40.56

6.27e–13

60.81

1.24e–14

50.34

3.14e–13

BiGRU + 3 Hidden

Layers NN

58.34

2.73e–15

45.63

1.42e–14

65.44

8.93e–16

55.85

4.53e–15

BiGRU + CNN

62.13

9.13e–16

50.28

3.24e–15

70.39

2.41e–17

60.57

1.82e–16

Back to article page