Page Areas:



Current Submenu:

Additional Information:

FAW @ Lange Nacht der Forschung 2016

Lange Nacht der Forschung 2016

Campus map

Campusplan JKU Linz

Our location on campus ...  more of Campus map (Titel)

Softwarepark Hagenberg

Hier den Alternativtext zum Bild eingeben!

Our location on Softwarepark Hagenberg ...  more of Softwarepark Hagenberg (Titel)


Position Indication:

Content

JoinVision

Online recruitment ("eRecruitment") enjoys increasing popularity. Nowadays, eRecruitment has become the primary source of recruitment processes, resulting in numerous job portals, trying to collect, centralize and summarize the large amount of job offers available on the Web. Web Information Extraction provides the basic technologies to overcome the manual collection of relevant data from Web sites. Screen scraping and rule-based tools exist, which can be adapted to the problem domain and thereafter allow the automatic extraction of key data.

The FFG–funded research project Jobolize is conducted by the Viennese job portal operator JoinVision E-Services in close cooperation with FAW. The aim of the project is the realization of a prototype demonstrating the effectiveness of new IE concepts, which were furtheron implemented in the IT-domain focused system joinvision.com. Companies are enabled to upload their IT-job offers to the JoinVision portal. Afterwards, they can search for fitting IT-professionals and get forwarded answers to their offers. IT-professionals, on the other hand, may enter their CVs, get informed on new job and project offers via e-mail and can actively search for jobs.

The claim of Jobolize is to automate JoinVision's acquiring process of job offers from Web pages by the means of semantic web and information extraction technologies.

In addition to the location and extraction of key data, e.g., job skills, language expertise, and job title, refining data, e.g., a skill level has to be determined.

The Jobolize prototype is based on the rule-driven text extraction tool GATE. By the means of a domain dependent "human resource" ontology, the semantic mapping of different skills and a rich named entity recognition is possible.

Furthermore, Jobolize cuts above the usual information extraction approaches by the means of a domain specific structure analysis approach. In detail, an identification of different areas of interest via page segmentation and elimination of irrelevant regions via site-level template detection is conducted. With this approach, a clear improvement of information extraction quality is possible.
-

Project Leader
Dipl.-Ing. Dr. Jürgen Palkoska
A.Univ.-Prof. DI Dr. Birgit Pröll

Project Staff
Mag. Simon Thalbauer
Dipl.-Ing. Christina Buttinger

Project website