Table 6. Aspect-context relation sentiment classification.
No | Model | SemI3 | AspR4 | PosI5 | SynI6 | Dataset | Acc (%) | Macro-F1 (%) |
---|---|---|---|---|---|---|---|---|
1 | ATAE-LSTM (Wang et al., 2016b) | ✓ | Res14 | 77.20 | – | |||
Lap14 | 68.70 | – | ||||||
2 | AF-LSTM (Tay, Tuan & Hui, 2018) | ✓ | Res14Res15 | 75.44 | – | |||
Lap14Lap15 | 68.81 | – | ||||||
3 | IAN (Ma et al., 2017) | ✓ | Res14 | 78.60 | – | |||
Lap14 | 72.10 | – | ||||||
4 | MGAN (Fan, Feng & Zhao, 2018) | ✓ | ✓ | ✓ | Res14 | 81.25 | 71.94 | |
Lap14 | 75.39 | 72.47 | ||||||
5 | ReMemNN (Liu & Shen, 2020) | ✓ | Res14 | 79.64 | 68.36 | |||
Lap14 | 71.58 | 65.41 | ||||||
6 | PF-CNN (Huang & Carley, 2018) | ✓ | Res14 | 79.20 | – | |||
Lap14 | 70.06 | – | ||||||
PG-CNN (Huang & Carley, 2018) | ✓ | Res14 | 78.93 | – | ||||
Lap14 | 69.12 | – | ||||||
7 | FDN (Shuang et al., 2020) | ✓ | Res14 | 82.30 | 75.00 | |||
Lap14 | 76.80 | 72.50 | ||||||
8 | Tensor DyMemNN (Tay, Tuan & Hui, 2017) | ✓ | Res14 | – | 58.61 | |||
Lap14 | – | 55.24 | ||||||
Holo DyMemNN (Tay, Tuan & Hui, 2017) | ✓ | Res14 | – | 58.82 | ||||
Lap14 | – | 60.11 | ||||||
9 | Co-attention+LRBP (Zhang et al., 2019) | ✓ | ✓ | Res14 | 80.35 | – | ||
Lap14 | 73.20 | – | ||||||
10 | MemNet (Tang, Qin & Liu, 2016b) | ✓ | ✓ | Res14 | 80.95 | – | ||
Lap14 | 72.21 | – | ||||||
11 | RAM (Chen et al., 2017b) | ✓ | ✓ | Res14 | 80.23 | 70.80 | ||
Lap14 | 74.49 | 71.35 | ||||||
12 | TNet-LF (Li et al., 2018) | ✓ | ✓ | Res14 | 80.79 | 70.84 | ||
Lap14 | 76.01 | 71.47 | ||||||
TNet-AS (Li et al., 2018) | ✓ | ✓ | Res14 | 80.69 | 71.27 | |||
Lap14 | 76.54 | 71.75 | ||||||
13 | AE-DLSTMs (Shuang et al., 2019) | ✓ | ✓ | Res14 | 79.57 | – | ||
Lap14 | 72.10 | – | ||||||
AELA-DLSTMs (Shuang et al., 2019) | ✓ | ✓ | Res14 | 80.35 | – | |||
Lap14 | 73.91 | – | ||||||
14 | History-based soft label model (Yin et al., 2019) | ✓ | ✓ | Res14 | 80.98 | 71.52 | ||
Lap14 | 74.56 | 71.63 | ||||||
15 | CapsNet (Jiang et al., 2019) | ✓ | 3*✓ | MAMS | 79.776 | – | ||
MAMS-small | 73.860 | – | ||||||
Res14 | 80.786 | – | ||||||
CapsNet-BERT (Jiang et al., 2019) | ✓ | ✓ | MAMS | 83.391 | – | |||
MAMS-small | 80.910 | – | ||||||
Res14 | 85.934 | – | ||||||
16 | IACapsNet (Du et al., 2019) | ✓ | ✓ | Res14 | 81.79 | 73.40 | ||
Lap14 | 76.80 | 73.29 | ||||||
17 | SAGAT (Huang et al., 2020) | ✓ | ✓ | Res14 | 85.08 | 77.94 | ||
Lap14 | 80.37 | 76.94 | ||||||
18 | ASGCN-DT (Zhang, Li & Song, 2019) | ✓ | ✓ | ✓ | Res14 | 80.86 | 72.19 | |
Res15 | 79.34 | 60.78 | ||||||
Res16 | 88.69 | 66.64 | ||||||
Lap14 | 74.14 | 69.24 | ||||||
ASGCN-DG (Zhang, Li & Song, 2019) | ✓ | ✓ | ✓ | Res14 | 80.77 | 72.02 | ||
Res15 | 79.89 | 61.89 | ||||||
Res16 | 88.99 | 67.48 | ||||||
Lap14 | 75.55 | 71.05 | ||||||
19 | CDT (Sun et al., 2019) | ✓ | ✓ | Lap14 | 77.19 | 72.99 | ||
Res14 | 82.30 | 74.02 | ||||||
Res16 | 85.58 | 69.93 |
Notes:
Semantic information.
Aspect-aspect relations.
Position information.
Syntax information.