Attention-based LSTM model with the Aspect information to solve financial opinion mining problem ( WWW 2018 shared task1 )
- python >=3.5
- tensorflow
- numpy
- pickle
- nltk
- gensim
run the script to finish preprocessing:
sh preprocess.sh
- aspect classification
python train_aspect.py
- sentiment analysis
python train_senti.py model_type # eg. python train_senti.py DeepMem, options are DeepMem or AT_LSTM
Shijia E. et al. Aspect-based Financial Sentiment Analysis with Deep Neural Networks.
@inproceedings{E.:2018:AFS:3184558.3191825,
author = {E., Shijia and Yang, Li and Zhang, Mohan and Xiang, Yang},
title = {Aspect-based Financial Sentiment Analysis with Deep Neural Networks},
booktitle = {Companion Proceedings of the The Web Conference 2018},
series = {WWW '18},
year = {2018},
isbn = {978-1-4503-5640-4},
location = {Lyon, France},
pages = {1951--1954},
numpages = {4},
url = {https://doi.org/10.1145/3184558.3191825},
doi = {10.1145/3184558.3191825},
acmid = {3191825},
publisher = {International World Wide Web Conferences Steering Committee},
address = {Republic and Canton of Geneva, Switzerland},
keywords = {long short-term memory network, representation learning, sentiment analysis},
}