An Approach Based on Multilevel Convolution for Sentence-Level Element Extraction of Legal Text

Table 3

The results on Divorce test dataset.

Models	HL (-)	Micro-P (+)	Micro-R (+)	Micro-F1 (+)
CNN	0.02516821	0.8624	0.8453	0.8537
CNN-RNN	0.02517237	0.8657	0.8532	0.8594
SGM	0.02197737	0.8810	0.8626	0.8717
SU4MLC	0.02497054	0.8796	0.8541	0.8667
ML-Reasoner	0.02631581	0.8901	0.8573	0.8733
Our model	0.02126052	0.8852	0.8728	0.8790