The goal of this thesis is to develop techniques to generate knowledge graph(s) (KG) by: recognizing and extracting entities and predicates from selected structural parts of transcriptions of our customer contacts - via chat and call - with our Customer Services center; mapping the extracted entities to our internal KPN business data model.
The created KGs will be queried in order to extract insights from customer contacts transcriptions and improve related business processes.
We plan to use Spacy and BERTje to perform POS, NER and triples extraction and store the labeled property graphs to either Neo4j or Tripply.