Leveraging the Inherent Hierarchy of Vacancy Titles for Automated Job\n Ontology Expansion

Machine learning plays an ever-bigger part in online recruitment, powering\nintelligent matchmaking and job recommendations across many of the world's\nlargest job platforms. However, the main text is rarely enough to fully\nunderstand a job posting: more often than not, much of the required information\nis condensed into the job title. Several organised efforts have been made to\nmap job titles onto a hand-made knowledge base as to provide this information,\nbut these only cover around 60\\% of online vacancies. We introduce a novel,\npurely data-driven approach towards the detection of new job titles. Our method\nis conceptually simple, extremely efficient and competitive with traditional\nNER-based approaches. Although the standalone application of our method does\nnot outperform a finetuned BERT model, it can be applied as a preprocessing\nstep as well, substantially boosting accuracy across several architectures.\n