Enhancement of Target-Oriented Opinion Words Extraction with Multiview-Trained Machine Reading Comprehension Model

Table 10

Performance on target extraction.

Models	14lap	14res	15res	16res	Average
RNCRF	74.92	75.18	74.14	73.12	74.34
CMLA	75.57	76.08	78.31	76.84	76.70
HAST	79.14	82.56	79.84	81.44	80.74
SpanMlt-BERT	80.78	84.26	77.71	80.95	80.92
MRC	80.31	86.73	77.85	81.59	81.62
MRC-MVT	82.64	87.15	78.42	83.24	82.86