¶ Entity Extraction
Entity Extraction is concerned with Entity Recognition and Entity Linking. This solution utilizes the SpaCy library to perform Entity Recognition, and the FuzzyWuzzy library for entity linking.
¶ External links
¶ GitHub
- The code is available at: https://github.com/Knox-AAU/PreProcessingLayer_EntityRecognitionAndLinking
¶ Overleaf 2023 Report
- The 2023 report is available at: https://www.overleaf.com/project/64feed8bda5b70b36afb6597
Every following section describes this pipeline in order, but first a visual overview.
¶ Overview
¶ How to get started
See the getting-started guide.
¶ The input that the solution takes
See the input explanation.
¶ Entity Recognition
Check out the entity-recognition documentation
¶ Entity Linking
Check out the entity-linking documentation
¶ The output it produces
See the output explanation.
¶ Deployment
See the docker-watchtower guide.
¶ Writing Tests
See the testing guide.
¶ Other components
- The directory-watcher
- The database
- The api
- The Language Detector
¶ Future work
The following points are elaborated in the 'Future Work' section of the Entity Extraction report
- Make a script that calculates the accuracy of the entity recognition and linker (ie. F1 Score)
- Train danish model to better recognize entities
- Make an entity linker with AI - Train it on the knowledge graph
- Proper exception handling
- Performance improvements
- Existing PostgreSQL database on the database server
- The storing of processed articles