A Sentence-Level Joint Relation Classification Model Based on Reinforcement Learning

Table 4

P@N comparison between different methods.

P@N (%)	100	200	300	Mean
LSTM	62	58	51.3	57.1
Bi-LSTM	73	65	59	65.6
Bi-LSTM-Attn	79	74	70.3	74.4
Bi-LSTM-Attn-RL	81	80	72.3	77.7
CNN	64	67	65.6	65.5
CNN-RL	78	76	70.6	74.8