Cross-Sectional Classification of Mental Health Risk Indicators in University Students Using Graph Convolutional Networks: The AtRisk-GCN Framework

Seifeddine Mechti1 Anis Saddoud2 Azza Allouch3 Salma Sassi4 Richard Chbeir5 Hajer Brahmi6 Boudour Ammar6 Mehdi Chlif3,7

  1. Faculty of Economic Science and managemement of Sfax, Larodec Laboratory, Tunis le Bardo, Tunis 3000, Tunisia seif.mechti@isims.usf.tn
  2. High Institute of Sport and Physical Education of Sfax, University of Sfax, Sfax 3000, Tunisia
  3. Faculty of Mathematical, Physical, and Natural Sciences of Tunis, University of El Manar, Tunisia / Robotics and Internet of Things Laboratory, Prince Sultan University, Riyadh, Saudi Arabia
  4. Department of Computer Engineering, Ecole Centrale d’Electronique, Lyon, 69007, France
  5. University Pau & Pays de l’Adour, E2S UPPA, LIUPPA, Anglet, 64600, France richard.chbeir@univ-pau.fr (corresponding author)
  6. REGIM-Lab: REsearch Groups in Intelligent Machines, National Engineering School of Sfax (ENIS), BP 1173, Sfax 3038, Tunisia Engineering School of Sfax (ENIS), Department of Management IT, Higher Business School of Sfax (ESCS), University of Sfax, Sfax 3038, Tunisia
  7. Research Laboratory: Sports Performance Optimisation, National Center of Medicine and Science in Sports (NCMSS), Tunis 263, Tunisia

Abstract

Identifying students at psychological risk is a growing public-health concern in higher education. Conventional self-report questionnaires are subject to response bias and provide only static snapshots of mental health status. Graph Convolutional Networks (GCNs) offer a principled way to model feature-based relational similarities between individuals, going beyond independent-sample as sumptions made by classical classifiers. In this paper we present AtRisk-GCN, a GCN-based framework for cross-sectional classification of questionnaire-derived and text-derived mental-health risk indicators in student populations. Two datasets are studied: (i) a survey-derived dataset of 1000 records, comprising real medical students from a single institution (Faculty of Biology and Medicine, University of Lausanne, and Lausanne University Hospital, CHUV, Switzerland), drawn from the publicly available data release accompanying Carrard et al. [5]) (Data 1, con tinuous psychological scale scores as node features) and (ii) a publicly available mental-health text corpus of 27977 samples (Data 2, TF-IDF representations as node features, no synthetic augmentation). In both cases, edges are constructed from cosine-similarity between node feature vectors using a k-nearest-neighbour (k-NN) scheme; no observed social-network data are available or assumed. On Data 1, AtRisk-GCN achieves an accuracy of 0.861 and an F1-score of 0.857, outperforming LSTM, CNN, and Transformer baselines. On Data 2 the model attains 0.84 accuracy and an F1-score of 0.84, surpassing, under matched input representation, TF-IDF based Logistic Regression and Linear SVM baselines, and clearly outperforming the graph-free MLP trained on identical features.

Key words

Graph Convolutional Networks, Mental Health Risk Classification, Student Well-being, Feature-similarity Graph, Node Classification, Cross-sectional Study, Data Augmentation Leakage.

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS260725053M