https://github.com/lixin4ever/hast

Aspect Term Extraction with History Attention and Selective Transformation (IJCAI 2018)
https://github.com/lixin4ever/hast

aspect-based-sentiment-analysis aspect-term-extraction opinion-target-extraction

Last synced: 2 months ago
JSON representation

Aspect Term Extraction with History Attention and Selective Transformation (IJCAI 2018)

• Host: GitHub
• URL: https://github.com/lixin4ever/hast
• Owner: lixin4ever
• Created: 2018-04-22T04:47:34.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2019-07-02T02:43:04.000Z (over 5 years ago)
• Last Synced: 2023-10-25T20:16:16.403Z (about 1 year ago)
• Topics: aspect-based-sentiment-analysis, aspect-term-extraction, opinion-target-extraction
• Language: Python
• Homepage: https://arxiv.org/abs/1805.00760
• Size: 1.28 MB
• Stars: 54
• Watchers: 7
• Forks: 11
• Open Issues: 3
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# HAST

Aspect Term Extraction with **H**istory **A**ttention and **S**elective **T**ransformation.

## Requirements

* Python 3.6

* [DyNet 2.0.2](https://github.com/clab/dynet) (For building DyNet and enabling the python bindings, please follow the instructions in this [link](http://dynet.readthedocs.io/en/latest/python.html#manual-installation))

* nltk 3.2.2

* numpy 1.13.3

## External Linguistic Resources

* [Glove Word Embeddings](https://nlp.stanford.edu/projects/glove/) (840B, 2.2M vocab).

* [MPQA Subjectivity Lexicon](http://mpqa.cs.pitt.edu/lexicons/subj_lexicon/)

## Preprocessing

* Window-based input (window size is 3, as done in Pengfei's [work](http://www.aclweb.org/anthology/D15-1168)).

* Replacing the punctuations with the same token `PUNCT`.

* Only the sentimental words with strong subjectivity are employed to provide distant supervision.

## Running

```

python main.py -ds_name [YOUR_DATASET_NAME] -sgd_lr [YOUR_LEARNING_RATE_FOR_SGD] -win [YOUR_WINDOW_SIZE] -optimizer [YOUR_OPTIMIZER] -rnn_type [LSTM|GRU] -attention_type [bilinear|concat]

```

## Environment

* OS: REHL Server 6.4 (Santiago)

* CPU: Intel Xeon CPU E5-2620 (Yes, we do not use GPU)

## Citation

If the code is used in your research, please star this repo and cite our paper as follows:

```

@inproceedings{li2018aspect,

  title={Aspect Term Extraction with History Attention and Selective Transformation},

  author={Li, Xin and Bing, Lidong and Li, Piji and Lam, Wai and Yang, Zhimou},

  booktitle={IJCAI},

  pages={4194--4200}

  year={2018}

}

```