This study addresses fairness in machine learning for student academic performance prediction using the UCI Student Performance dataset. We comparatively evaluate logistic regression, Random Forest, and XGBoost, integrating the Synthetic Minority Oversampling Technique (SMOTE) to address class imbalance and 5-fold cross-validation for robust model training. A comprehensive fairness analysis is conducted, considering sensitive attributes such as gender, school type, and socioeconomic factors, including parental education (Medu and Fedu), cohabitation status (Pstatus), and family size (famsize). Using the AIF360 library, we compute the demographic parity difference (DP) and Equalized Odds Difference (EO) to assess model biases across diverse subgroups. Our results demonstrate that XGBoost achieves high predictive performance (accuracy: 0.789; F1 score: 0.803) while maintaining low bias for socioeconomic attributes, offering a balanced approach to fairness and performance. A sensitivity analysis of bias mitigation strategies further enhances the study, advancing equitable artificial intelligence in education by incorporating socially relevant factors.

The increasing demand for privacy-preserving, ethically aligned synthetic data generation in education has highlighted the limitations of existing tabular data generators. Traditional approaches often sacrifice fairness or privacy in pursuit of predictive accuracy, rendering them unsuitable for high-stakes academic settings. In this paper, we propose FairSYN-Edu, a novel diffusion-based synthetic data generation framework designed for educational data. By integrating adversarial debiasing and differentially private training into the generative process, FairSYN-Edu jointly optimizes utility, fairness, and privacy. We evaluate our approach on three real-world educational datasets spanning MOOC, K–12 tutoring, and LMS environments. Experimental results demonstrate that FairSYN-Edu achieves significantly lower demographic disparities, maintains competitive predictive performance (RMSE = 0.402), and provides moderate resistance to membership inference attacks (AUC = 0.705). Ablation studies, error gap analysis, and SHAP-based interpretability evaluations confirm the robustness and ethical reliability of our method. We release the complete implementation, synthetic benchmark suite, and documentation to promote reproducibility and responsible AI practices in education.

Bu derleme çalışması, yapay zekâ tabanlı karar destek sistemlerinin (YZ-KDS) teknik, etik ve yönetişim boyutlarını sistematik bir analiz çerçevesinde incelemeyi amaçlamaktadır. Sağlık, finans, üretim/lojistik ve çevre yönetimi gibi kritik sektörlerde YZ-KDS sistemlerinin uygulamaları, veri kalitesi, açıklanabilirlik, adalet ve hesap verebilirlik bağlamında ele alınmıştır. Çalışmanın veri temeli, Kesgin ve Dözer (2025) tarafından geliştirilen BiBLoX yazılımı aracılığıyla Web of Science, Scopus ve TR Dizin veri tabanlarından çekilen 4824 bibliyografik kayda dayanmaktadır. PRISMA protokolü doğrultusunda yapılan filtreleme sonucunda 1273 yayın detaylı incelemeye alınmıştır. Bibliyometrik analiz sonuçlarına göre, açıklanabilir yapay zekâ çerçeveleri (XAI), federated learning tabanlı gizlilik çözümleri ve adil karar mekanizmaları, sistemlerin güvenilirliğini artırmada öne çıkan yaklaşımlar olarak belirlenmiştir. Ayrıca disiplinlerarası iş birliği, etik denetim ve kurumsal yönetişim süreçlerinin, bu sistemlerin sorumlu ve sürdürülebilir biçimde geliştirilmesi açısından kritik olduğu vurgulanmaktadır. Çalışma, yapay zekâ destekli karar destek sistemlerinin gelecekteki tasarımlarına hem teorik hem uygulamalı katkı sunmayı hedeflemektedir.

Artificial intelligence (AI) is rapidly transforming science education by enabling adaptive learning, virtual experimentation, and generative applications such as ChatGPT. Yet, research in this area remains fragmented and often descriptive. This study adopts an integrated approach, combining bibliometric analysis, meta-analysis, and predictive modeling, to map the research landscape, synthesize empirical impacts, and forecast future directions of AI in science education. The bibliometric analysis of 3217 studies (2015–2024) identified four major thematic clusters: intelligent tutoring and adaptive learning, virtual and remote laboratories, generative AI and ethics, and teacher professional development. The meta-analysis of 39 empirical studies revealed a significant overall positive effect of AI on science learning outcomes (Hedges’ g = 0.63, \(I^2 = 89\%\)), with stronger effects in higher education (g = 0.75) and low- and middle-income country (LMIC) contexts (g = 0.82). Predictive modeling using Support Vector Regression (SVR) projected continued growth in AI and science education research until 2030, with signs of thematic saturation after 2028. The findings extend constructivist, sociocultural, and Bloom’s frameworks by showing how AI fosters individualized inquiry, collaborative learning, and higher-order cognitive tasks. Practically, they highlight the need for teacher training frameworks such as Technological Pedagogical Content Knowledge (TPACK) and the 5E model to ensure meaningful integration of AI tools. Policy implications include the development of AI literacy curricula, certification standards for educational AI applications, and ethical governance guidelines for data use and privacy. This study contributes uniquely by moving beyond descriptive reviews, offering predictive foresight that informs educators, policymakers, and researchers about both current trends and future trajectories of AI in science education.

The accelerating pace of scientific production has outpaced the capabilities of traditional bibliometric tools, which have laid the groundwork for quantifying scholarly impact and mapping collaboration networks, enabling significant advances in the field. While these methods have significantly shaped the field, their reliance on static, retrospective analyses limits their ability to navigate the hyperdynamic and interdisciplinary research ecosystem. In response, we propose BiBLoX: a pioneering, predictive bibliometric intelligence platform that builds on these foundations by integrating real-time data streams exclusively via official APIs from sources such as Web of Science, Scopus, and TRDizin, accessed through institutional subscriptions and in compliance with their terms of service. This ensures legally compliant, ethically sound, and technically robust data acquisition without any reliance on web scraping, alongside machine learning-driven trend forecasting and dynamic knowledge mapping. Evaluations across 123,638 publications, validated through temporal cross-validation and robustness testing, show BiBLoX’s ability to predict citation trajectories, identify emerging topics, and visualize evolving collaboration networks. Compared to existing platforms, BiBLoX delivers predictive analytics with reproducible datasets and rigorous validation, paving the way for proactive scientific foresight. By providing stakeholders with dynamic, actionable insights, BiBLoX enables evidence-based funding decisions, innovation policy design, and strategic positioning within rapidly evolving knowledge economies.

This study conducts a bibliometric analysis of Intelligent Transportation Systems (ITS) research, using data from Web of Science (WoS), Scopus, and TR Dizin covering 2015–2024. The analysis reveals a 440% increase in annual publications, driven by advancements in artificial intelligence (AI), the Internet of Things (IoT), and sustainability-focused technologies. Dominant research themes include IoT, machine learning, and smart city solutions, with sustainability-related topics such as eco-routing and renewable energy integration gaining prominence. Geographic analysis identifies India, China, and the United States as leading contributors, while emerging economies like Turkey and South Korea are expanding their research footprints. Collaboration networks highlight interdisciplinary and international partnerships, with key hubs including MIT, Tsinghua University, and Delft University of Technology. Machine learning models predict steady growth in ITS publications, projecting approximately 950 annual outputs by 2026. Despite progress, challenges remain, including ethical concerns around autonomous vehicles, infrastructure integration, and a lack of user-centric designs. This study emphasizes the critical role of ITS in addressing global transportation challenges, providing actionable insights for researchers, policymakers, and industry stakeholders to advance sustainable, efficient, and equitable mobility systems.