HUAPA is the proposed model in 《Improving Review Representations with User Attention and Product Attention for Sentiment Classification》, which is accepted by AAAI'18.
The original datasets are released by the paper [Tang et al., 2015]. [Download]
The embedding file. [Download]
For example, you can use the folowing command to train HUAPA in the dataset imdb:
python train.py --n_class 10 --dataset imdb
The best model will be saved in the folder "../checkpoints/imdb/timestamp". The timestamp is server time.
For example, you can use the folowing command to test HUAPA in the dataset imdb:
python test.py --n_class 10 --dataset imdb --checkpoint ../checkpoints/imdb/timestamp
if you use the code, please cite the following paper:
[Wu et al., 2018] Zhen Wu, Xin-Yu Dai, Cunyan Yin, Shujian Huang, Jiajun Chen. Improving Review Representations with User Attention and Product Attention for Sentiment Classification. In Proceedings of AAAI.
[Wu et al., 2018] Zhen Wu, Xin-Yu Dai, Cunyan Yin, Shujian Huang, Jiajun Chen. Improving Review Representations with User Attention and Product Attention for Sentiment Classification. In Proceedings of AAAI.
[Tang et al., 2015] Duyu Tang, Bing Qin, Ting Liu. Learning Semantic Representations of Users and Products for Document Level Sentiment Classification. In Proceedings of EMNLP.