Neural Network based model for Aspect-Based Sentiment Analysis.
- NOTE: it is NOT related to our finished or ongoing research projects.
- Word embeddings: stanford GloVe
- Ctx Feat Extractor: CNN + Multi-Channel
- Target Feat Extractor: Weighted sum of word vectors making up the target phrase
| 14semval-restaurant | 14semeval-laptop | ||
|---|---|---|---|
| ATAE-LSTM [1] | 77.2/- | 68.7/ | - |
| MemNet [2] | 78.16/65.83 | 70.33/64.09 | 68.50/66.91 |
| IAN [3] | 78.6/- | 72.1/- | - |
| RAM [4] | 80.23/70.80 | 74.49/71.35 | 69.36/67.30 |
| Model 1 | 79.43/69.49 | 74.65/69.27 | 71.10/69.32 |
- Attention-based LSTM for Aspect-level Sentiment Classification. EMNLP 2016
- Aspect Level Sentiment Classification with Deep Memory Network. EMNLP 2016
- Interactive Attention Networks for Aspect-Level Sentiment Classification. IJCAI 2017
- Recurrent Attention Network on Memory for Aspect Sentiment Analysis. EMNLP 2017
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent