stripathi08/hline

Hierarchical Language Identifier and Named Entity Recognizer

FIRE 2015 : Mixed Script Information Retrieval

Shared Task Description at this link.

Subtask 1

All the code is in this repository.

To test the code:

1. Add the training and testing datasets in the corresponding folders.
2. Ready to use models, are on this link. Just unzip and store in the Resources/ folder. You can then run main_demo.py to check a demo of the classifiers at work. You can run main.py to generate a submission without the NE's subclassified.
3. NER subclassification will require you to add your bing api key at helpers.py and running the code ne_subcategory_classification.py once the submission has been generated to generate a new submission with NE's subclassified.

We will soon release a library tool for ease of usage, so check this page for future updates