PPP-QuestionParsing-ML
This project is a complete reformulation of the Python 3 project PPP-QuestionParsing-ML-standalone. This work is currently in progress. Right now, two functions are implemented:
- feature extraction (build a vector with an english sentence).
- Loading of the dataset
Dependencies:
- StanfordParser: this project uses directly the StanfordParser, so you have to download the sources of StanfordParser and to link them in your java IDE to the project.
- A look up table:
curl http://metaoptimize.s3.amazonaws.com/cw-embeddings-ACL2010/embeddings-scaled.EMBEDDING_SIZE=25.txt.gz | gunzip > embeddings-scaled.EMBEDDING_SIZE=25.txt