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

Table 4

The results on Loan dataset test dataset.

Models	HL (-)	Micro-P (+)	Micro-R (+)	Micro-F1 (+)
CNN	0.03033127	0.8418	0.8397	0.8407
CNN-RNN	0.03015758	0.8369	0.8406	0.8387
SGM	0.02958177	0.8432	0.8370	0.8401
SU4MLC	0.02883519	0.8555	0.8227	0.8388
ML-Reasoner	0.01861665	0.8719	0.8317	0.8513
Our model	0.02772556	0.8633	0.8573	0.8603