Aspect-Based Sentiment Analysis on Laptop14 original (test)
85.74AccuracyMT-ISA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MT-ISAModel=Flan-T5 XXL (11B), Strategy=Output Strategy (O)2024.12 | 85.74 | 83.86 | |
| MT-ISAModel=Flan-T5 XXL (11B), Strategy=Input Strategy (I)2024.12 | 85.4 | 82.73 | |
| MT-ISAModel=Flan-T5 Base (250M), Strategy=Input Strategy (I)2024.12 | 82.91 | 79.86 | |
| BERT Asp + SCAPT†Model=BERT-Base (110M)2024.12 | 82.76 | 79.15 | |
| ABSA-ESA*Model=T5-Base (220M)2024.12 | 82.44 | 79.34 | |
| THORModel=Flan-T5 XXL (11B)2024.12 | 82.29 | 78.78 | |
| Direct fine-tuneModel=Flan-T5 Base (250M)2024.12 | 81.98 | 77.93 | |
| MT-ISAModel=Flan-T5 Base (250M), Strategy=Output Strategy (O)2024.12 | 81.98 | 78.1 | |
| Direct fine-tuneModel=Flan-T5 XXL (11B)2024.12 | 81.82 | 77.69 | |
| BERT Asp + CEPT†Model=BERT-Base (110M)2024.12 | 81.66 | 78.38 | |
| THORModel=Flan-T5 Base (250M)2024.12 | 81.66 | 77.51 | |
| C³DA+Model=BERT-Base (110M)2024.12 | 80.61 | 77.11 | |
| ISAIVModel=BERT-Base (110M)2024.12 | 80.41 | 77.25 | |
| RGATModel=BERT-Large (340M)2024.12 | 80.25 | 76.26 | |
| GPT-4o-miniProtocol=Zero-shot2024.12 | 79.94 | 73.83 | |
| InstructABSAModel=TK-INSTRUCT (200M)2024.12 | 79.31 | 46.29 | |
| BERT-ADA+Model=BERT-Base (110M)2024.12 | 78.96 | 74.18 | |
| BERT-SPC+Model=BERT-Base (110M)2024.12 | 78.22 | 73.45 | |
| RGATModel=BERT-Base (110M)2024.12 | 78.21 | 74.07 | |
| GPT-4o-mini + THORProtocol=Zero-shot2024.12 | 72.1 | 55.42 |