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

Table 11

Performance on pair extraction.

Models	14lap	14res	15res	16res	Average
HAST + IOG	53.41	62.39	58.12	63.84	59.44
JERE-MHS	52.34	66.02	59.64	58.58	59.14
SpanMlt-BERT	65.75	72.72	61.06	69.58	67.27
HAST + MRC	64.71	73.13	67.92	73.68	69.86
MRC-MVT	67.35	77.02	68.63	75.99	72.25