Table 8 Paired samples t-test results comparing SBJLIS with baseline models on the ATIS dataset in terms of F1-score.

From: Siamese-based metric joint learning for intent detection and slot filling using triplet loss optimization

Baseline model

Mean Diff

Std. Dev

95% CI (Lower–Upper)

t

df

p

Slot-Gated

2.580

0.021

[2.553, 2.606]

271.96

4

 < 0.001

Bi-directional

2.710

0.036

[2.666, 2.754]

171.40

4

 < 0.001

CTRAN

3.410

0.036

[3.366, 3.453]

215.67

4

 < 0.001

CEA-Net

2.800

0.050

[2.739, 2.861]

126.49

4

 < 0.001

Back to article page