Transformer-based Multi-Aspect Modeling for Multi-Aspect Multi-Sentiment Analysis. Zhen Wu, Chengcan Ying, Xinyu Dai, Shujian Huang, Jiajun Chen. NLPCC 2020.
- pytorch=1.7.1
- python=3.7
Download the pretrained RoBERTa model from Baiduyun (password:2fv2) or Google Drive and unzip it into the folder pretrained.
The MAMS data is preprocessed by the script preprocess.py. The original and preprocessed versions of the data are provided in the folder data.
Run the command python main_ATSA.py to train and test the ATSA model.
You can change training settings in the file configs.py.
Run the command python main_ACSA.py to train and test the ACSA model.
You can change training settings in the file configs.py.
If you use the code, please cite our paper:
@InProceedings{10.1007/978-3-030-60457-8_45,
author="Wu, Zhen
and Ying, Chengcan
and Dai, Xinyu
and Huang, Shujian
and Chen, Jiajun",
editor="Zhu, Xiaodan
and Zhang, Min
and Hong, Yu
and He, Ruifang",
title="Transformer-Based Multi-aspect Modeling for Multi-aspect Multi-sentiment Analysis",
booktitle="Natural Language Processing and Chinese Computing",
year="2020",
publisher="Springer International Publishing",
address="Cham",
pages="546--557",
isbn="978-3-030-60457-8"
}[1]. Zhen Wu, Chengcan Ying, Xinyu Dai, Shujian Huang, Jiajun Chen. Transformer-based Multi-Aspect Modeling for Multi-Aspect Multi-Sentiment Analysis. NLPCC 2020.
[2]. Qingnan Jiang, Lei Chen, Ruifeng Xu, Xiang Ao, Min Yang. A Challenge Dataset and Effective Models for Aspect-Based Sentiment Analysis. EMNLP-IJCNLP 2019.