Dataset and source code for our paper: Inateraction Attention Transfer Network for Cross-domain Sentiment Classification.
The public dataset has been uploaded.
Crowdfunding project Dataset: Indiegogo.com
This is our private dataset, if you want to use it, please indicate the source, thank you！
— Python 2.7.5
—Tensorflow-gpu 1.2.1
— Numpy 1.13.3
— sklearn
— other pakages
—To install requirements, please run pip install -r requirements.txt.
— OS: CentOS Linux release 7.7.1908
— CPU: 24 E5-2650 v4 @ 2.20GHz
— GPU: 4 * K80:11441 MB
Prepare the Pre-trained Word2vec :
— 1. Get the pre-trained model and generate the embeddings ;
— Google Word2Vec ;
— GloVe ;
— 2. Put the pre-trained word_embedding (Google-Word2Vec/Glove) to the coresponseding path ;
Prepare the aspect sequence :
— python aspect_extraction.py
(Input/output path can be changed inner the file!)
Run the model :
— python train.py
(Default dataset is Laptop; The parameters can be changed in the train.py file! (line 15~line 31))
If you have any problem about this library, please send us an Email at:
— kkzhang0808@mail.ustc.edu.cn
If the data and code are useful for your research, please be kindly to give us stars and cite our paper as follows:
@article{zhang2019interactive,\
title={Interactive Attention Transfer Network for Cross-domain Sentiment Classification},\
author={Zhang, Kai and Zhang, Hefu and Liu, Qi and Zhao, Hongke and Zhu, Hengshu and Chen, Enhong},\
year={2019}\
}