I'm working on creating a knowledge graph of relationships between entities in emails, using neo4j. I now need to convert natural language search queries into cypher queries. I have been unable to find anything concrete with regards to this. How do I proceed?
1 Answers
Great paper on this topic: https://dspace.mit.edu/bitstream/handle/1721.1/119708/1078222310-MIT.pdf?sequence=1
This involves, at the core:
- Named Entity Recognition (custom NER with SpaCy, https://towardsdatascience.com/custom-named-entity-recognition-using-spacy-7140ebbb3718)
- Named Entity Linking (with a Knowledge Base, https://medium.com/analytics-vidhya/entity-linking-a-primary-nlp-task-for-information-extraction-22f9d4b90aa8)
- Relation Extraction (https://medium.com/@andreasherman/different-ways-of-doing-relation-extraction-from-text-7362b4c3169e) (although, the RE method in the aforementioned paper seems more concrete)
I'm not an expert on this at all- just beginning to learn about NLP, graph databases, etc.
So if anyone has additional resources or notes on the ones I shared, please contribute!
Related