Implementation of Nested Named Entity Recognition
Some files are part of NeuroNLP2.
Requirements
We tested this library with the following libraries:
Running experiments
Testing this library with a sample data
Run the gen_data.py to generate the processed data files for training, and they will be placed at the "./data/" directory
python gen_data.py
Run the train.py to start training
python train.py
Reproducing our experiment on the ACE-2004 dataset
Put the corpus ACE-2004 into the "../ACE2004/" directory
Put this .tgz file into the "../" and extract it
Run the parse_ace2004.py to extract sentences for training, and they will be placed at the "./data/ace2004/"
python parse_ace2004.py
Run the gen_data_for_ace2004.py to prepare the processed data files for training, and they will be placed at the "./data/" directory
python gen_data_for_ace2004.py
Run the train.py to start training
python train.py
Reproducing our experiment on the ACE-2005 dataset
Put the corpus ACE-2005 into the "../ACE2005/" directory
Put this .tgz file into the "../" and extract it
Run the parse_ace2005.py to extract sentences for training, and they will be placed at the "./data/ace2005/"
python parse_ace2005.py
Run the gen_data_for_ace2005.py to prepare the processed data files for training, and they will be placed at the "./data/" directory
python gen_data_for_ace2005.py
Run the train.py to start training
python train.py
Reproducing our experiment on the GENIA dataset
Put the corpus GENIA into the "../GENIA/" directory
Run the parse_genia.py to extract sentences for training, and they will be placed at the "./data/genia/"
python parse_genia.py
Run the gen_data_for_genia.py to prepare the processed data files for training, and they will be placed at the "./data/" directory
python gen_data_for_genia.py
Run the train.py to start training
python train.py
Configuration
Configurations of the model and training are in config.py
Citation
Please cite our arXiv paper:
@article{shibuya2019nested,
title={Nested Named Entity Recognition via Second-best Sequence Learning and Decoding},
author={Shibuya, Takashi and Hovy, Eduard},
journal={arXiv preprint arXiv:1909.02250},
year={2019}
}