Effat University develops AI model to predict student performance and improve academic support
Researchers at Effat University have developed a machine learning-based software tool that predicts students' future academic performance.
Dhekra Alkaf, Faigah Bajammal, and Manal Asrar designed the system to support academic advisors in identifying students who may require additional support, enabling institutions to intervene early and allocate resources more effectively.
The model, which does not require a background in computer science to operate, is designed for ease of use by academic counselors and faculty. By analyzing undergraduate students' GPAs from their first three semesters, along with other academic and demographic data, the system predicts final semester GPAs with a high degree of accuracy.
How the model works
The AI-driven tool processes multiple data points, including:
Academic records – Assignment scores, quizzes, class tests, and attendance records.
Demographic factors – Age, gender, family background, and special needs.
Researchers encountered challenges in educational data mining, particularly in ensuring data privacy and ethical collection practices. Future research will explore ways to refine data handling and improve prediction accuracy.
Machine learning approach and results
The study found that the Random Forest algorithm was the most effective method for predicting student GPAs.
A Random Forest algorithm is a machine learning technique that combines multiple decision tree algorithms to improve prediction accuracy. Decision trees work by splitting data into smaller groups based on specific variables (such as attendance or assignment scores) to determine likely outcomes. However, decision trees on their own can be prone to overfitting, meaning they might make predictions too closely tied to the training data rather than identifying broader patterns.
By averaging the results of multiple decision trees, Random Forest reduces errors and increases accuracy, making it particularly effective for predicting student performance.
Compared to logistic regression and artificial neural networks, the Random Forest model demonstrated superior performance.
Logistic regression is a statistical method used to predict binary outcomes (such as pass/fail). While useful for classification tasks, it does not perform as well when dealing with complex, multi-variable data like student performance.
Artificial neural networks (ANNs) are designed to mimic the way the human brain processes information, using layers of interconnected nodes to detect patterns in data. While powerful, ANNs require large datasets and significant computational resources, making them less practical for this specific application.
The study found that Random Forest outperformed both logistic regression and artificial neural networks, providing institutions with a reliable method for identifying students who may need academic support.
According to the research, applying machine learning to student data could help universities optimize learning outcomes and intervention strategies. The study states, “Through the application of machine learning on educational data through prediction, universities and educational institutions will be able to improve teaching and learning outcomes, as well as provide the right support for the different types of students at the institution.”
Potential benefits for higher education
The predictive model could support universities in:
Identifying students at risk and providing targeted academic interventions.
Improving resource allocation by directing support to students who need it most.
Reducing dropout rates and increasing graduation rates by addressing challenges earlier.
Aligning with global education initiatives, such as the United Nations Sustainable Development Goal for inclusive and equitable education.
Ethical considerations and challenges
Despite its potential, the technology raises concerns regarding:
Data privacy – Ensuring that student records are securely stored and accessed responsibly.
Bias in algorithms – Preventing unintentional discrimination against certain student demographics.
Accuracy of predictions – Avoiding misclassification of students who may not need intervention.
Misuse of data – Ensuring the technology is used to support students rather than limit resources.
Researchers emphasize the need for responsible AI use to maximize benefits while addressing ethical challenges.
