yanshengli / ABSC

aspect-based sentiment classification

Aspect-based Sentiment Classification

Introduction

This is the Tensorflow source code of our paper

Two-side Target Representation with Target-Context Rotatory Attention for Aspect-based Sentiment Analysis. Submitted to CIKM 2018.

Meanwhile, we provide our implementations of some state-of-the-art ABSC models.

If you use this package, please cite our paper.

Related Papers

1. Duyu Tang, Bing Qin, Xiaocheng Feng, and Ting Liu. Effective LSTMs for Target-Dependent Sentiment Classification with Long Short Term Memory. COLING 2016.

2. Yequan Wang, Minlie Huang, Li Zhao, and Xiaoyan Zhu. Attention-based LSTM for Aspect-level Sentiment Classification. EMNLP 2016.

3. Duyu Tang, Bing Qin, and Ting Liu. Aspect Level Sentiment Classification with Deep Memory Network. EMNLP 2016.

4. Meishan Zhang, Yue Zhang, and Duy-Tin Vo. Gated Neural Networks for Targeted Sentiment Analysis. AAAI 2016.

5. Dehong Ma, Sujian Li, Xiaodong Zhang, and Houfeng Wang. Interactive Attention Networks for Aspect-Level Sentiment Classification. IJCAI 2017.

6. Peng Chen, Zhongqian Sun, Lidong Bing, and Wei Yang. Recurrent Attention Network on Memory for Aspect Sentiment Analysis. EMNLP 2017.

7. Two-side Target Representation with Target-Context Rotatory Attention for Aspect-based Sentiment Analysis. Submitted to CIKM 2018.

source code tree

.
├── README.md
├── model
│   ├── lstm.py          Paper 1
│   ├── tc_lstm.py       Paper 1
│   ├── td_lstm.py       Paper 1
│   ├── at_lstm.py       Paper 2
│   ├── dmn_lstm.py      Paper 3
│   ├── ian.py           Paper 5
│   ├── ram.py           Paper 6
│   ├── lcr.py           Paper 7

Usage

Usage of codes:

Usage: python model/lcr.py  [options]   [parameters]
Options:
        --train_file_path
        --test_file_path
        --embedding_file_path
        --learning_rate
        --batch_size
        --n_iter
        --random_base
        --l2_reg
        --keep_prob1
        --keep_prob2

Give the usage of lcr.py for example:

python model/lcr.py --train_file_path data/absa/laptop/laptop_2014_train.txt
                    --test_file_path data/absa/laptop/laptop_2014_test.txt
                    --embedding_file_path data/absa/laptop/laptop_word_embedding_42b.txt
                    --learning_rate 0.1
                    --batch_size 25
                    --n_iter 50
                    --random_base 0.1
                    --l2_reg 0.00001
                    --keep_prob1 0.5
                    --keep_prob2 0.5