Complementary Peer Matching Using Vector Space Modeling and K-Means Clustering in Intelligent E-Learning Systems

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος:	Complementary Peer Matching Using Vector Space Modeling and K-Means Clustering in Intelligent E-Learning Systems
Συγγραφείς:	Thomas, Eapen, Joseph, Binumon
Στοιχεία εκδότη:	Zenodo, 2026.
Έτος έκδοσης:	2026
Θεματικοί όροι:	Peer Matching, K-Means Clustering, Vector Space Modeling, Cosine Similarity, E-Learning, Study Buddy, Learner Profiling, LearnSphere, Collaborative Learning, FastAPI, Supabase
Περιγραφή:	Finding a proper study partner is a major dilemma to students who have to learn in an online environment. Unlike in physical classroom where socialization among the students goes almost naturally, e-learning environments rarely provide systematic means of linking students based on complementary academic profiles. This description of the Study Buddy functionality which has been incorporated into LearnSphere, an AIenhanced Learning Management System, handles this problem through an algorithmic peer-matchmaking pipeline. The system is functioning on two layers. It first builds normalised skill vectors based on quiz and assignment performance of each student, then ranks pairs of candidates based on a weighted expression that captures both joint curriculum coverage and diversity of profiles between two students as measure that is cosine-distanced. The second layer uses a KMeans clustering model on data on realtime platforms to assign students to three learner archetypes with a small bonus extended to cross-cluster pairings. Once they have successfully matched, the students are able to dispatch buddy requests, chat, and/or schedule study sessions, all on the same platform.
Τύπος εγγράφου:	Article Conference object
DOI:	10.5281/zenodo.19543944
DOI:	10.5281/zenodo.19543945
Rights:	CC BY
Αριθμός Καταχώρησης:	edsair.doi.dedup.....c9a4e07d1999a343da657207c2b8426e
Βάση Δεδομένων:	OpenAIRE

Περιγραφή
DOI:	10.5281/zenodo.19543944