Automatic Generation of the Draft Procuratorial Suggestions Based on an Extractive Summarization Method: BERTSLCA

Table 2

ROGUE1 comparisons between different methods.

Method\value	P	R	F1
LEAD	19.3	14.4	16.5
NEURALSUM	20.5	30.1	24.4
TextRank	21.2	26.3	23.5
BERTSUM + FC	21.8	31.5	25.8
BERTSUM + Transformer	28.2	26.8	27.5
BERTSUM + LSTM	27.6	24.9	26.2
BERTSUM + BiLSTM + attention	23.1	33.6	27.2
BERTSUM + CNN	50.1	18.4	26.9
BERTSUM + BiLSTM + CNN	50.8	19.0	27.4
Proposed method	51.0	19.3	28.1

Note. The value of the proposed method is the average value after 10-fold cross-validation test.