Finding the perfect job that takes into account someone's skills and ambitions is an overwhelming challenge for many freshly graduated students. At the same time, companies struggle to hire employees fulfilling their requirements. Ideally, a successful match of students and jobs includes the preferences of both sides. This paper proposes a reciprocal recommender system matching graduating students and job offers. To construct a common representation space for the items of the recommendation, the course descriptions from the curriculum are used as a linking factor. Further, Latent Dirichlet Allocation (LDA) is used to extract topics from the course and job descriptions, forming the latent representation space. Next to providing job recommendations, curricula gaps can be discovered and thus the students can be better prepared for a future career. The algorithm is tested on data from a Data Science bachelor programme. Results show that a reciprocal matching of graduating students and jobs generates recommendations with precision and recall up to 0.7. The approach yields promising results for such system to be implemented as a job recommender system on university level or as a stand-alone system for graduating students.
|Title of host publication
|Proceedings of the 12th International Conference on Educational Data Mining, EDM 2019, Montréal, Canada, July 2-5, 2019
|Michel C. Desmarais, Collin F. Lynch, Agathe Merceron, Roger Nkambou
|International Educational Data Mining Society
|Published - Jul 2019