A Reciprocal Recommender System for Graduates' Recruitment

A reciprocal recommender system for graduates' recruitment (RRSGR) is presented in this paper. RRSGR makes full use of the historical information in university about graduates and former graduates. Both collaborative filtering and content-based filtering are incorporated in RRSGR, which help to achieve a win-win situation between graduates and employers. Probabilistic neighborhood selection and priority k-medoids clustering are adopted to improve the diversity of recommendation results. A preliminary evaluation has been conducted based on the real data of Shandong Normal University. Experiments and user survey show that RRSGR is effective to improve the accuracy and diversity of recruitment recommendation.