A Markov Chain Model for Learning-State Transitions in Outcome-based Education
Year: 2026
Type: article
Abstract: This study presents a cross-sectional management framework for classifying students' academic performance in the outcome-based education (OBE) context using publicly available school-level academic performance data. The framework incorporates academic and engagement-related indicators, including study time, absences, previous academic failures, and final academic achievement, to classify students into high-performing, satisfactory, and at-risk categories. The analysis uses the mathematics subset of the Student Performance dataset, comprising 395 student records. Under the adopted classification rules, 15 students (3.80%) are classified as high-performing, 243 (61.52%) as satisfactory, and 137 (34.68%) as at-risk. Correlation analysis indicates that previous failures have a negative association with final grade (r=-0.360), whereas study time and absences have weak positive associations (r=0.098 and r=0.034, respectively). A sensitivity analysis further shows that the category proportions vary with changes in the classification thresholds. The proposed framework provides a mathematically interpretable and educationally meaningful descriptive approach for identifying groups of students who may require further academic monitoring and support in an OBE environment. Because the dataset contains one record per student, the study does not estimate longitudinal transitions, stationary distributions, or the effects of interventions. The findings are therefore interpreted as cross-sectional results rather than evidence of student progression over time.
Authors: Akbota Nurgaliyeva, Angisin Seitmuratov
Institution: Korkyt Ata Kyzylorda State University
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7233605161
FWCI:
Cites:
Cited by:
Related to:
