/knowledge-graph

Primary LanguageJupyter Notebook

Information Extraction (Entities & Relations) & Knowledge Graph Construction from Raw Text

This repositoy contains the code for the construction of Domain KG from Unstructired text data by extracting the domain information (entities and relation) using NLP techniques.

  • Extraction of Entities & Relation using rule-based approach (AllenNLP, NLTK, SpaCy, Stanford CoreNLP)
  • Coreference Resolution (neuralcoref, allen nlp)
  • Knowledge Graph Construction based on Domain Ontology (Neo4j)
  • Contextual Graph embedding generation using Neo4j DataScience library (Node2Vec, GraphSAGE) to train Neural Collaborative Filtering (NCF) model for recommendation.
  • Template-based Query segment generation from User Natural Query.