AI-based identification and support of at-risk students: A case study of the Moroccan education system
Project Overview
The document explores the application of generative AI in education, focusing on a predictive modeling approach designed to identify at-risk students within the Moroccan education system. By employing machine learning techniques and analyzing historical data, the framework aims to significantly reduce dropout rates through timely interventions. The methodology demonstrates high accuracy and precision in predicting students' risks, highlighting the necessity of comprehensive data analysis and the collaboration between educators and data scientists. This adaptable approach underscores the potential of AI to enhance educational outcomes across various educational contexts, fostering a proactive response to student needs and ultimately contributing to improved retention and success rates in schools.
Key Applications
AI-driven predictive modeling for identifying at-risk students
Context: Moroccan education system targeting students at risk of dropout
Implementation: Utilized machine learning techniques on historical educational data provided by the Moroccan Ministry of National Education
Outcomes: Achieved 88% accuracy, 88% recall, 86% precision, and an AUC of 87%; enabled timely interventions for at-risk students
Challenges: Data imbalance and prediction accuracy in identifying the exact moment of dropout
Implementation Barriers
Data-related barrier
Imbalanced datasets where dropout cases are significantly fewer than non-dropout cases, affecting model training and performance.
Proposed Solutions: Employ resampling techniques, class weighting, and ensemble models to improve predictions across both classes.
Implementation barrier
Difficulty in pinpointing the exact moment of dropout due to the unpredictable nature of student behavior and external factors.
Proposed Solutions: Incorporate survival analysis techniques to better predict the timing of dropouts.
Project Team
Ismail Elbouknify
Researcher
Ismail Berrada
Researcher
Loubna Mekouar
Researcher
Youssef Iraqi
Researcher
El Houcine Bergou
Researcher
Hind Belhabib
Researcher
Younes Nail
Researcher
Souhail Wardi
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Ismail Elbouknify, Ismail Berrada, Loubna Mekouar, Youssef Iraqi, El Houcine Bergou, Hind Belhabib, Younes Nail, Souhail Wardi
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai