A Hybrid Neural Network BERT-Cap Based on Pre-Trained Language Model and Capsule Network for User Intent Classification

Table 6

The performance comparison of different methods using F1 score measure.

Methods	F1 score
CNN + domain template	0.899
Lib-SVM + n-gram	0.912
CNN + rules	0.926
CNN + ensemble learning	0.929
LSTM + domain dictionary	0.941
LR + character + POS tags + target words + semantic tags	0.945
BERT-Cap + focal loss	0.967