Analysis and Comparison of Unsupervised Algorithms to Detect Student Dropout Risk
Keywords:
k-means, fuzzy c-means, education, school dropout, classification, artificial intelligence.Abstract
The COVID-19 pandemic had significant repercussions across various sectors, with the education sector being one of the most affected. The inclusion, continuity, and timely graduation of students enrolled in higher education are among the priorities established in the General Education Law. To achieve this, strategies and measures must be implemented to promote student retention in higher education institutions. For this reason, one of the primary educational challenges is providing teachers with the necessary tools and resources to identify and refer cases involving violence, mental health concerns, and potential school dropout.
Through the application of data mining techniques in education, it has been possible to predict academic performance, create predictive models for student retention, and define behavioral profiles. A review of the literature has shown that student dropout is influenced by multiple factors, such as academic, economic, and social variables. By utilizing clustering algorithms, more detailed information and insights into dropout patterns can be obtained, supporting informed decision-making in higher education.
In this project, the indicators will be analyzed and defined to monitor the academic performance of students at the ITESG campus in two phases: first using all student features and then using only the relevant features to compare results. This dataset will be managed by an intelligent system that will use unsupervised artificial intelligence algorithms, such as k-means and fuzzy c-means (FCM), to cluster students and detect potential dropout cases, which can then be referred to the tutoring area. The results suggest that the FCM algorithm performs best for detecting students at risk of dropping out.
The objective of this work is to analyze and compare the performance of the unsupervised algorithms for detecting students at risk of dropping out, using a dataset from the Instituto Tecnológico Superior de Guanajuato and considering both the full set of features and only the most relevant ones identified through statistical tests.
References
[1] Subsecretaría de Educación Superior, "Criterios Generales para la Distribución de los Recursos Autorizados al Programa Presupuestario U079", Programa de Expansión de la Educación Media Superior y Superior, 2023.
[2] Secretaria de Educación Pública, "Programa Sectorial de Educación 2020-2024", 2020.
[3] A. Urbina Nájera, "Deserción escolar universitaria, Patrones para prevenirla aplicando minería de datos educativa", e-Journal of Educational Research, Assessment and Evaluation, 2020.
[4] H. Sadiq, "Educational data mining and analysis of students’ academic performance using WEKA", Indonesian Journal of Electrical Engineering and Computer Science, vol. 9, no. 2, pp. 447-459, 2018.
[5] F. Guiyun, "Analysis and prediction of students’ academic performance based on educational data mining", IEEE Access, vol. 10, pp. 19558-19571, 2022.
[6] A. Ani, "Enhancing the clustering of student performance using the variation in confidence", Intelligent Tutoring Systems: 14th International Conference, pp. 274-279, 2018.
[7] R. González, "Massive LMS log data analysis for the early prediction of course-agnostic student performance", Computers & Education, vol. 163, 2021.
[8] F. I. Moreira, "Moodle Predicta: A Data Mining Tool for Student Follow Up", CSEDU, pp. 339-346, 2017.
[9] M. Saqr, "How learning analytics can early predict under-achieving students in a blended medical education course", Medical Teacher, pp. 757-767, 2017.
[10] D. Y. Hooshyar, "Clustering algorithms in an educational context: An automatic comparative approach", IEEE Access, vol. 8, pp. 146994-147014, 2018.
[11] Y. G. Li, "Educational data mining for students' performance based on fuzzy C‐means clustering", The Journal of Engineering, pp. 8245-8250, 2019.
[12] M. Jessica, "DETECT: a hierarchical clustering algorithm for behavioural trends in temporal educational data", Artificial Intelligence in Education: 21st International Conference, pp. 374-385, 2020.
[13] B.-A. Norka, "Classification Models for Determining Types of Academic Risk and Predicting Dropout in University Students", International Journal of Advanced Computer Science and Applications(IJACSA), vol. 11, 2020.
[14] B. Andreas, "Early Prediction of University Dropouts – A Random Forest Approach", Jahrbücher für Nationalökonomie und Statistik, vol. 240, no. 6, pp. 743-789, 2020.
[15] K. Lorenz, "Predicting student dropout: A machine learning approach", European Journal of Higher Education, 2020.
[16] V. Matti, "Predicting student dropouts with machine learning: An empirical study in Finnish higher education", Technology in Society, vol. 76, 2024.
[17] S. Martín, "Perspectives to Predict Dropout in University Students with Machine Learning", IEEE International Work Conference on Bioinspired Intelligence (IWOBI), 2018.
[18] H. Sadiq, "Classification, clustering and association rule mining in educational datasets using data mining tools: A case study", Cybernetics and Algorithms in Intelligent Systems: Proceedings of 7th Computer Science On-line Conference 2018, pp. 196-211, 2019.
[19] B. J.C., "Pattern Recognition with Fuzzy Objective Function", Springer, 1981.
[20] R. &. S. R. Suganya, "Fuzzy c-means algorithm-a review", International Journal of Scientific and Research Publications, vol. 2, no. 11, 2012.
[21] T. J., "Feature selection for classification: A review", Data classification: Algorithms and applications, 2014.
