Syntax-Based Skill Extractor for Job Advertisements
Ellery Smith, Martín Braschler, Andreas Weiler, Thomas Haberthuer · 2019
In the context of extracting relevant skill-terms from job advertisements, we propose a syntax-based method for generating large amounts of machine-labelled text from a small amount of human-labelled data. This is then used to solve the vocabulary problem and significantly increase recall when detecting skills.