JobOlize - Headhunting by Information Extraction in the Era of Web 2.0

E-recruitment is one of the most successful ebusiness applications supporting both, headhunters and job seekers. The explosive growth of online job offers makes the usage of information extraction techniques to build up, e.g., job portals in a semiautomatic way a necessity. Existing approaches, however, hardly cope with the heterogeneous and semistructured nature of job offers nor do they consider potentials offered by Web 2.0 technologies. This paper proposes an information extraction system called “JobOlize”1, realized for arbitrarily structured IT job offers. To improve extraction quality, a hybrid approach is employed, combining existing NLPtechniques with a new form of context-driven extraction, incorporating layout, structure and content information. To allow users a proper adaptation of the extraction results while preserving the look and feel of the original Web pages, a rich client interface is provided. The improvements in extraction quality are justified on basis of a case study and the experiences gained are generalized and critically reflected by discussing lessons learned.