Con2KG - A large-scale domain-specific knowledge graph

Abstract

This paper presents Con2KG, a large-scale recruitment domain Knowledge Graph that describes 4 million triples as facts from 250 thousands of unstructured data of job postings. We propose a novel framework for Knowledge Graph construction from unstructured text and an unsupervised, dynamically evolving ontology that helps Con2KG to capture hierarchical links between the entities missed by explicit relational facts in the triples. To enrich our graph, we include entity context and its polarity. Towards this end, we discuss Con2KG applications that may benefit the recruitment domain.