spaCy Performance on Named Entity Recognition with Code-Mixed Data

Project work for seminar Computational Modeling's BN requirements.

About: This project aims to examine multi-functional spaCy language models' performance on named entity recognition (NER) tasks, when the texts are code-mixed.

Repository Structure

• /data: instructions on how to get the dataset used in the experiment
• /notebooks: intermediate development workspaces for the project
• /scripts: implementation of the project
• /text_compose: organizational texts and their source LaTeX codes

How2Run

• Install runtime requirements in REQUIREMENTS.txt.
• Follow the instructions in /data/How2GetData.txt to download the correct dataset.
• Make sure train.conll is in the /data folder.
• Run notebook /scripts/spaCyNEROnCS.py.

License

All source code is made available under the GPL-3.0 license. You can freely use and modify the code without warranty if you provide attribution to the authors. See LICENSE for the full license text.

All contents under /text_compose are not open source. The author reserves the rights to the content.