Sentiment Classification on Yelp 2013
67.8AccuracyCHIM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CHIMBase Model=BiLSTM, Injection=embedding2019.08 | 67.8 | 0.646 | |
| CHIMBase Model=BiLSTM, Injection=classifier2019.08 | 67.5 | 0.641 | |
| BiLSTM + linear-basis-custBackbone=BiLSTM, Customization=linear-basis-cust2019.02 | 67.1 | 0.662 | |
| CHIMBase Model=BiLSTM, Injection=encoder2019.08 | 67 | 0.659 | |
| BiLSTM + bias-basis-custBackbone=BiLSTM, Customization=bias-basis-cust2019.02 | 66.9 | 0.654 | |
| CHIMBase Model=BiLSTM, Injection=attention2019.08 | 66.5 | 0.664 | |
| CMAArchitecture=HierAttention + attention-cust (CMA)2019.02 | 66.4 | 0.677 | |
| CMABase Model=LSTM+HierAtt, Injection=attention, separate bias-attention for user and product=true2019.08 | 66.4 | 0.677 | |
| DUPMNArchitecture=HierLSTM + memory-cust2019.02 | 66.2 | 0.667 | |
| DUPMNBase Model=HierLSTM, Injection=memory networks2019.08 | 66.2 | 0.667 | |
| BiLSTM + encoder-basis-custBackbone=BiLSTM, Customization=encoder-basis-cust2019.02 | 66.1 | 0.665 | |
| PMAArchitecture=HierLSTM + attention-cust (PMA)2019.02 | 65.8 | 0.668 | |
| PMABase Model=HierLSTM, Injection=attention, uses additional external features=true2019.08 | 65.8 | 0.668 | |
| HCSCArchitecture=BiLSTM + CNN + attention-cust (CSAA)2019.02 | 65.7 | 0.66 | |
| HCSCBase Model=BiLSTM+CNN, Injection=attention, considers cold-start entities=true2019.08 | 65.7 | 0.66 | |
| NSCArchitecture=LSTM + attention-cust2019.02 | 65 | 0.692 | |
| NSCBase Model=HierLSTM, Injection=attention2019.08 | 65 | 0.692 | |
| UPDMNArchitecture=LSTM + memory-cust2019.02 | 63.9 | 0.662 | |
| UPDMNBase Model=LSTM, Injection=memory networks2019.08 | 63.9 | 0.662 | |
| UPNNArchitecture=CNN + word-cust + bias-cust2019.02 | 59.6 | 0.784 | |
| UPNNBase Model=CNN, Injection=embedding, classifier2019.08 | 59.6 | 0.748 |