Work

A decision-aware framework for pre-regulatory academic early warning under resource constraints

Year: 2026

Type: article

Abstract: Early-warning systems in higher education are usually evaluated as prediction tools, but institutional value depends on how predictive information is converted into timely support decisions. We evaluated a decision-aware machine-learning framework for pre-regulatory academic warning using longitudinal administrative records from six undergraduate cohorts at a regional public university. The framework estimated future violation of formal academic-warning conditions and converted calibrated risks into support decisions through context-sensitive expected violation costs. Compared with regulation-only, rule-based, probability-threshold, and cost-insensitive predictive policies, the proposed decision policy reduced missed early intervention under matched resource constraints and narrowed disparities in missed intervention across heterogeneous learning-condition groups. These findings suggest that academic early-warning systems should be assessed not only by discrimination or calibration, but also by decision quality, resource efficiency, and equity of institutional support. Decision-aware learning analytics can help universities act before formal warning thresholds are breached while preserving human oversight and regulatory accountability.

Source: Scientific Reports

DOI: 10.1038/s41598-026-72395-0

Authors: Trung-Nghia Phung, Gia-Nhu Nguyen, Dac-Nhuong Le, The-Vinh Nguyen

Institutions: Thai Nguyen University of Information and Communication Technology, Duy Tan University, Hai Phong University

Language: English

Topic: Online Learning and Analytics

Subfield: Computer Science Applications

Field: Computer Science

Domain: Physical Sciences

SDG: Quality education

Open Access status: gold

Estimated APC paid: $2,690