Table 5. Classification performance of models on English datasets (where AC represents aspect category and SP represents sentiment polarity).

Bold and underlined values indicate the best result.

Model	AC				SP
	IMDB	Twitter	Lap14	Rest16	IMDB	Twitter	Lap14	Rest16
	Acc. Macro-F1	Acc. Macro-F1	Acc.Macro-F1	Acc. Macro-F1	Acc. F1	Acc. F1	Acc. F1	Acc. F1
ATAE-LSTM	69.21 68.24	69.65 67.40	69.14 63.18	83.25 63.85	77.20 68.70	76.60 68.90	70.80 69.00	71.50 69.50
RAM	70.23 70.80	69.36 67.30	74.49 71.35	85.5865.76	70.80 80.23	74.49 71.35	69.36 73.85	72.14 60.24
IAN	73.72 69.21	60.49 67.33	74.31 58.28	85.44 65.99	78.60 72.10	76.70 68.29	76.70 76.70	72.31 65.35
PBAN	84.60 76.60	78.74 72.61	87.8174.12	81.67 73.68	81.16 74.12	78.7570.51	78.6670.84	78.0470.74
BiGCN	81.32 71.72	72.20 70.45	75.6370.74	87.71 67.87	82.93 75.29	76.27 72.39	75.04 73.85	74.59 72.86
ASGCN	80.77 72.02	72.15 70.40	75.55 71.05	88.9967.48	71.92 70.63	73.51 68.83	79.40 69.43	79.40 61.18
DualGCN	83.1073.01	73.29 72.02	75.53 72.01	86.24 67.62	84.27 78.08	78.84 74.74	75.92 74.29	74.45 69.37
KumaGCN	81.97 73.38	74.16 73.35	74.59 71.84	88.96 70.84	89.3973.19	89.1270.89	89.2372.04	88.64 73.82
PT-BERT	80.98 72.04	73.82 69.35	72.10 69.90	81.16 73.20	73.98 69.94	72.11 70.74	81.33 73.57	74.96 70.93
BERT-Pair	87.9 79.8	92.78 89.07	93.30 90.89	86.90 83.70	97.00 93.70	93.57 90.83	95.60 92.18	89.90 85.90
RoBERTa	81.42 71.79	75.36 71.11	73.78 72.37	82.76 75.25	77.43 74.21	75.43 74.04	81.60 72.48	75.96 71.96
CG-BERT	89.1079.70	94.38 90.97	95.60 92.64	90.10 86.80	97.2093.60	94.2790.12	95.8092.14	90.40 86.90