Ranking-Based At-Risk Student Prediction Using Federated Learning and Differential Features

2025-05-14

AI Summary

To address the challenge of cross-institutional academic risk prediction under educational data privacy constraints, this paper proposes a ranking-oriented predictive framework integrating differential feature engineering with federated learning. Unlike conventional absolute-value modeling, we introduce a novel relative-change-based differential feature representation that preserves privacy without sharing raw log or grade data. The framework employs a ranking-centric multi-objective evaluation—incorporating Top-n accuracy, normalized discounted cumulative gain (nDCG), and precision-recall area under the curve (PR-AUC)—to enhance early-risk detection capability. Trained on 12 courses involving 1,136 students and evaluated on 5 unseen courses, our method achieves ranking performance comparable to centralized baselines while significantly outperforming non-differential federated baselines. It enables accurate identification of at-risk students as early as the beginning of the semester.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Data transparency and provenance

Abstract

Digital textbooks are widely used in various educational contexts, such as university courses and online lectures. Such textbooks yield learning log data that have been used in numerous educational data mining (EDM) studies for student behavior analysis and performance prediction. However, these studies have faced challenges in integrating confidential data, such as academic records and learning logs, across schools due to privacy concerns. Consequently, analyses are often conducted with data limited to a single school, which makes developing high-performing and generalizable models difficult. This study proposes a method that combines federated learning and differential features to address these issues. Federated learning enables model training without centralizing data, thereby preserving student privacy. Differential features, which utilize relative values instead of absolute values, enhance model performance and generalizability. To evaluate the proposed method, a model for predicting at-risk students was trained using data from 1,136 students across 12 courses conducted over 4 years, and validated on hold-out test data from 5 other courses. Experimental results demonstrated that the proposed method addresses privacy concerns while achieving performance comparable to that of models trained via centralized learning in terms of Top-n precision, nDCG, and PR-AUC. Furthermore, using differential features improved prediction performance across all evaluation datasets compared to non-differential approaches. The trained models were also applicable for early prediction, achieving high performance in detecting at-risk students in earlier stages of the semester within the validation datasets.

Problem

• Predict at-risk students across schools while preserving privacy
• Enhance model generalizability using differential features in EDM
• Combine federated learning and relative values for early prediction

Innovation

• Federated learning ensures privacy-preserving model training
• Differential features enhance model performance and generalizability
• Ranking-based prediction for at-risk students using multi-course data