Just-In-Time Defect Prediction (JITDP), as an important means to improve software quality and reduce maintenance costs, has received widespread attention in recent years. However, existing methods generally neglect multi-scale temporal features during the development process, lack dynamic modeling of developer behavior and project lifecycle, and have limited robustness when facing concept drift. To address these limitations, this paper proposes a novel Multi-Scale Temporal Defect Prediction framework (MSTDP) that integrates commit behavior patterns at temporal granularities of hours, days, and weeks, combining code semantic information, developer behavioral features, and lifecycle-aware mechanisms to comprehensively characterize the dynamic evolution patterns of defect introduction. MSTDP achieves deep feature fusion through cross-attention mechanisms and constructs a concept drift analysis module to enhance the model’s adaptability across different evolutionary stages. Empirical results on 9 large-scale open-source projects demonstrate that MSTDP achieves optimal performance across all evaluation metrics: average accuracy of 76.89% (0.56% improvement), precision of 61.02% (0.53% improvement), recall of 67.82% (0.64% improvement), F1-score of 60.93% (0.46% improvement), and AUC of 81.88% (0.52% improvement), significantly outperforming 7 mainstream baseline methods including JITBoost, FENSE, and JIT-CF. Wilcoxon signed-rank tests confirm the statistical significance of these improvements (\(\varvec{p < 0.05}\)). The research results validate the effectiveness of this method for efficient and accurate defect prediction in dynamic software engineering environments, demonstrating stronger generalization capability and robustness against concept drift.
6 August 2025 in Journal of Social Systems and Policy Analysis
Student dropout prediction is a critical challenge in higher education that requires accurate identification of at-risk students to enable timely interventions. This study presents EASE-Predict (Ensemble-SHAP Explainable Student Prediction), a comprehensive ensemble learning framework with SHAP-based explainable AI to predict student academic outcomes. We evaluated five machine learning algorithms (Random Forest, Gradient Boosting, Extra Trees, Logistic Regression, and SVM) and developed voting and stacking ensemble models on a dataset of 4,424 students with 36 features encompassing academic performance, socioeconomic factors, and demographic information.EASE-Predictachieved superior performance with 77.4% accuracy, representing a statistically significant improvement of 4.3 percentage points over the best individual model (Random Forest: 77.3%). The framework demonstrated exceptional class-specific discriminative performance with AUC scores of 0.930 for Graduate prediction (vs. 0.927 for best individual model), 0.821 for Enrolled students (vs. 0.794 for SVM), and 0.913 for Dropout identification (vs. 0.904 for individual models). Cross-validation results showed superior stability with the lowest performance variance (σ = 0.014 vs. σ = 0.0189 for Random Forest). SHAP explainability analysis quantified feature importance, revealing that second semester curricular units completion accounts for 60% of prediction influence, followed by tuition payment status (35%) and scholarship availability (12%).McNemar’s statistical tests confirmed that EASE-Predict’s performance improvements are statistically significant (p < 0.05) across all evaluation metrics.The framework maintains interpretability while achieving state-of-the-art accuracy, providing educational institutions with actionable insights for implementing evidence-based intervention strategies.
Stock price modeling under privacy constraints presents a unique challenge at the intersection of computational economics and machine learning. Financial institutions and brokerage firms hold valuable yet sensitive data that cannot be centrally aggregated due to privacy laws and competitive concerns. To address this issue, we propose a novel Fast-Converging Federated Learning (FCFL) framework that enables decentralized and privacy-preserving stock price modeling. FCFL employs a dual-stage adaptive optimization strategy that dynamically tunes local learning rates and aggregation weights based on inter-client gradient divergence, accelerating convergence in heterogeneous financial environments. The framework integrates secure aggregation and differential privacy mechanisms to prevent information leakage during communication while maintaining model fidelity. Experimental results on multi-institutional stock datasets demonstrate that FCFL achieves up to 30% faster convergence and 2.5% lower prediction error compared to conventional federated averaging approaches, while guaranteeing strong -differential privacy. Theoretical analysis further proves that the framework attains sublinear convergence in communication rounds under non-IID data distributions. This study provides a new direction for collaborative financial modeling, balancing efficiency, accuracy, and privacy in real-world economic systems.
27 March 2026 in Environmental Science & Technology