<![CDATA[Beyond identifying students at risk of academic burnout, higher education institutions also need an evidence-based mechanism to segment the student population and allocate limited wellbeing-support resources proportionally. This study proposes a decision support framework combining K-Means clustering with the CRiteria Importance Through Intercriteria Correlation (CRITIC) objective weighting method and the COmplex PRoportional ASsessment (COPRAS) ranking method to segment students and prioritize wellbeing-support allocation. Using a dataset of 60 students with eight actionable psychosocial and behavioral indicators, K-Means partitioned students into three risk-based segments (High/Moderate/Low), evaluated via silhouette scores (0.133-0.163 for k = 2-5) and set at k = 3 to align with the institution's three-tier intervention scheme. CRITIC derived data-driven, correlation-based criteria weights without expert elicitation, unlike preference-based weighting. These weights fed into COPRAS to compute a relative significance score (Qi) and rank students by support-allocation priority. The High-risk cluster captured four of five dataset-labelled High-burnout students (80%) and showed the highest mean Burnout Risk Score (41.60) versus Moderate-risk (37.03) and Low-risk (30.09) clusters. The COPRAS ranking correlated significantly with independent burnout indicators (Spearman's rho = 0.706 with Burnout Risk Score, 0.749 with Support Need, and 0.529 with Intervention Urgency, all p < 0.001), with mean ranks decreasing monotonically across Burnout_Level categories (High = 7.0, Moderate = 23.3, Low = 43.1, out of 60). These findings indicate that unsupervised segmentation combined with objective multi-criteria weighting can provide a transparent, replicable, resource-efficient basis for allocating student wellbeing support, complementing supervised prediction-based approaches used in prior studies.]]>
Copyrights © 2025
