Modeling Student Task Group Preferences Using Graph Theory and Spectral Clustering

Abstract

This study aims to model student preferences in forming task groups using graph theory and clustering algorithms. The research object involves 2024 cohort students of the IF A Siang class at Universitas Satya Terra Bhinneka. Preference data were collected through questionnaires and transformed into numerical representations for analysis. Graph theory was applied to model relationships between students based on preference similarity, while spectral clustering was used to form optimal student groups. The results show that spectral clustering is able to identify groups with high internal similarity and clear separation between clusters. This approach provides an objective alternative to conventional group formation methods and helps minimize dissatisfaction among students. The proposed model can support lecturers in forming balanced and effective task groups based on student preferences.