A Study Partner Recommender System Using a Community Detection Algorithm

The potential for an individual’s interaction to be limited by an underlying inclination has significant implications when overlooked in situations where group contact is crucial. In university course projects, the absence of synergy among a project team often leads to failure. In order to provide students enrolled in our course projects with an optimal team recommendation, we have devised a highly efficient strategy. This strategy involves utilizing their personalities and assessing their collaborative effectiveness through individual personality questionnaires. Additionally, we employ community detection using the Leiden algorithm. In order to assess the soundness of our recommendation technique, we conducted an evaluation of current algorithms that have been employed for this purpose, taking into account their modularity scores and the decision to establish preferred team sizes. The assessment section presents results that demonstrate how the Leiden algorithm outperforms competing algorithms and techniques that rely on clustering coefficients. Specifically, the Leiden algorithm is capable of recommending study partners that possess many personality features that match the user and their respective teams.