NJUNLP / TOWE

Code and data for "Target-oriented Opinion Words Extraction with Target-fused Neural Sequence Labeling" (NAACL2019)

Target-oriented Opinion Words Extraction with Target-fused Neural Sequence Labeling

This repository is under construction.

Dependency:

• python3
• pytorch 0.4

How to run (Take the 14res dataset as example)

1. (optional) prepare the word embeddings: (we have prepared for you in the code/data/ directory.)
   1. put the glove embeddings in the code/embedding directory.
   2. run the script:

   python build_vocab_embed.py --ds 14res
   

2. in the code/ directory:

   python main.py --ds 14res