Table 8 Significance analysis of the models’ performance with different feature extractor on ChiNesE dataset.
From: Transformer-based prototype network for Chinese nested named entity recognition
Feature Extractor | Bert | RoBerta | longformer | roformer | -BiLSTM |
|---|---|---|---|---|---|
\(\mu \left( \% \right)\) | 84.64 | 84.11 | 84.09 | 76.47 | 84.27 |
\(\sigma\) | 0.026 | 0.018 | 0.031 | 0.103 | 0.023 |
n | 5 | 5 | 5 | 5 | 5 |
\(\rho\) | - | 0.052 | -0.665 | -0.190 | -0.133 |
\(\Delta\) | - | 0 | 0 | 0 | 0 |
df | - | 4 | 4 | 4 | 4 |
\(t \, Stat\) | - | -5.80 | -3.96 | -47.29 | -3.48 |
\(P(T<t)\) | - | 0.002 | 0.008 | 0.001 | 0.013 |
t | - | 2.132 | 2.132 | 2.132 | 2.132 |
