Skip to content

Latest commit

 

History

History
73 lines (55 loc) · 2.56 KB

readme.md

File metadata and controls

73 lines (55 loc) · 2.56 KB

ABSC_APWEBWAIM

This is an aspect-based sentiment analysis project. We implemented many models(LSTM, ATAE-LSTM, BILSTM-ATT-G, GCAE, Memnet, RAM, TNet). If you use this code, please cite the original paper and this project.

Requirements

  • python 3.6
  • pytorch 1.0
  • tensorboardX
  • tqdm
  • nltk
  • absl-py

Usage

# prepro
python -m lstm.main --mode prepro
# train
python -m lstm.main --mode train
# test
python -m lstm.main --mode test
# You can set different parameters or use different models and datasets.

Experiment Results

Models Restaurant Laptop Twitter
LSTM 76.70/63.57 69.28/63.30 66.04/63.46
ATAE-LSTM 77.23/63.73 69.44/63.46 71.24/69.19
GCAE 78.12/62.50 70.38/64.02 72.40/70.89
BILSTM-ATT-G 79.20/67.07 71.32/64.88 71.68/70.37
Memnet 77.86/64.47 68.18/62.46 69.80/66.86
RAM 78.30/65.42 71.63/66.73 71.24/68.75
TNet 78.39/65.37 73.98/68.64 72.11/70.01

Models

ATAE-LSTM Wang, Yequan, Minlie Huang, and Li Zhao. "Attention-based LSTM for aspect-level sentiment classification." Proceedings of the 2016 conference on empirical methods in natural language processing. 2016.

GCAE Xue, Wei, and Tao Li. "Aspect based sentiment analysis with gated convolutional networks." arXiv preprint arXiv:1805.07043 (2018).

BILSTM-ATT-G Liu, Jiangming, and Yue Zhang. "Attention modeling for targeted sentiment." Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers. 2017.

Memnet Tang, Duyu, Bing Qin, and Ting Liu. "Aspect level sentiment classification with deep memory network." arXiv preprint arXiv:1605.08900 (2016).

RAM Chen, Peng, et al. "Recurrent attention network on memory for aspect sentiment analysis." Proceedings of the 2017 conference on empirical methods in natural language processing. 2017.

TNet Li, Xin, et al. "Transformation networks for target-oriented sentiment classification." arXiv preprint arXiv:1805.01086 (2018).