Multigranularity Pruning Model for Subject Recognition Task under Knowledge Base Question Answering When General Models Fail

Table 2

Experimental results for overall accuracies (%).

Method	Accuracy (%)
MemNN-ensemble [48]	63.9
CFO [1]	75.7
BiLSTM-CRF + BiLSTM [55]	78.1
Structure attention + MLTA [5]	82.3
KEQA [56]	75.4
M3M [53]	76.9
BERT-CRF (baseline)	82.6
EGP [52]	81.4
SSMFRP [57]	83.0
BERT-CRF + MGPM	86.0

Bold values represent the best-performance.