About

We at Semantic Web Company see ourselves as a learning organization. As such, we want to make the best out of the strengths and skills of our employees. It is important to us that our employees can have meaningful exchanges with their coworkers, find projects that match their interests and develop themselves by finding open positions within our organization. The HR Recommender is a semantic matchmaking tool that supports this effort by providing a visualized dashboard and sorted lists. 

Based on your profile information it connects you with other employees, shows you interesting projects, and open positions within your organization. With its innovative and powerful knowledge model, the HR Recommender extracts keywords from your profile to generate a semantically enriched footprint that describes your skills, interests and strengths.

Improve your Footprint

For quality matches and to ensure you get great results, be sure to create your own profile by uploading your CV, adding keywords as well as writing about yourself.

Fine-tune your matches

Get the most out of the HR Recommender by adjusting the sliders in the recommender system so that your strengths influence the search more or less.

Knowledge model

The core functionality of the HR Recommender comes from a powerful knowledge model, containing skills, topics, and occupations. The knowledge model has been built with PoolParty and contains over 21,000 concepts.

The HR Recommender uses an adapted and enhanced version of the ESCO classification of the European Commission as a basis for skills and the occupations that are related to them.

Method

The matchmaking process:

  1. A semantic footprint is generated from the user profile data. The footprint contains concepts that have been extracted from the HR Recommender knowledge graph. The data and the footprint are stored in Drupal and the PoolParty GraphSearch index.
  2. In order to find suitable matches, the footprint is enriched with further information from the knowledge graph. Not only are exact matches from the footprint used, but also related concepts, siblings, broader and narrower concepts. Most importantly skills are enriched with relevant job positions on the basis of the ESCO classification, and vice versa. Each type of match is scored differently, which is then used to calculate the overall score of an extracted concept. The result of this process is a vector with the extracted concepts and their scores.
  3. This vector is used to find matching employees, projects and open positions in our Solr index. Solr calculates the scores of the matches (the closer the match, the higher the score) and sorts them. The result is a sorted list of employees, projects, and open positions.
  4. The result list is then displayed in the HR Recommender app in Drupal.

 

Team

Kurt Moser, Juliane Pineiro-Winkler, Andreas Blumauer