Knowledge Discovery for Higher Education Student Retention Based on Data Mining: Machine Learning Algorithms and Case...

Area of Science:

• Data Science
• Educational Data Mining
• Machine Learning

Background:

• Student retention is a critical issue affecting higher education efficiency.
• Predicting student dropout is essential for implementing timely interventions.
• Existing studies have not fully explored the impact of socioeconomic factors on retention.

Purpose of the Study:

• To develop and validate machine learning models for predicting student retention at different academic levels.
• To identify key variables influencing student dropout in higher education.
• To provide actionable insights for institutions to reduce dropout rates.

Main Methods:

• Employed data mining and machine learning algorithms including decision trees, k-nearest neighbors, logistic regression, naive Bayes, random forest, and support vector machines.
• Utilized higher education data to build predictive models for student retention across three academic years.
• Applied class balancing techniques to improve algorithm performance.

Main Results:

• Achieved prediction accuracy exceeding 80% across all retention levels.
• Identified secondary educational score and community poverty index as significant predictive variables for student dropout.
• Random forest algorithm demonstrated superior performance among the tested machine learning techniques.

Conclusions:

• Machine learning models can effectively predict student retention in higher education.
• Socioeconomic factors play a significant role in student dropout, necessitating targeted interventions.
• Institutions can leverage predictive analytics to proactively support students and improve retention rates.