Jul 2026· 2026 4th International Conference on Sustainable Computing and Smart Systems (ICSCSS)· pp. 1861-1867· 0 citations· 15 references
Abstract
In India, engineering college admissions continue to be a high stakes affair, especially when being applied in the Tamil Nadu Engineering Admissions (TNEA) system where students of varied socioeconomic backgrounds are subjected to obscured decision-making processes and dearth of guidance. The current paper describes a machine learning system that is both data-driven and fairness-conscious to increase the degree of transparency and fairness in the process of admission. The suggested system has a hybrid architecture which consists of regression-based cutoff prediction engine with a fairness-constrained recommendation engine which matches students to appropriate colleges based on academic and contextual profiles. Ensemble regressors are applied to historical data, such as cut off scores, group codes, the type of college, and counseling marks, yielding a predictive accuracy of 92.6% (R2 score) and mean absolute error (MAE) of 4.12. Such measures of fairness as demographic parity difference (DPD) and equal opportunity difference (EOD) are less than 0.03 between caste and gender groups, which means that there is a small amount of bias. The hybrid recommender combines both content-based and collaborative filtering and has a top-5 recommendation accuracy of 94.6% and Mean Reciprocal Rank (MRR) of 0.837. A mobile/web interface provides real-time and personalized advice to students and families. This system does not only enhance the results of the admissions but also creates an algorithmic transparency and social inclusivity in counseling processes. The architecture is ethically aligned, scalable and modular that it can be deployed in various states. The findings confirm the validity, equity and practicality of the system. Further expansions of it will incorporate reinforcement learning and real-time feedback loops to improve recommendation accuracy and equity adherence.
The use of machine learning in recruitment has raised growing concerns about fairness, as automated hiring systems can generate unequal outcomes across demographic groups. These disparities are influenced not only by imbalanced data but also by the behavior of learning algorithms, making bias a multidimensional challenge that cannot be effectively addressed with single-stage solutions. This study introduces an integrated framework for bias detection and mitigation in AI-driven recruitment systems, combining interventions at the data, model, and decision levels within a unified evaluation pipeline. The framework is assessed using multiple classification models of varying complexity and evaluated with established fairness metrics. In addition, explainability techniques are employed using SHAP-based feature attribution to investigate hidden dependencies and assess the sensitivity of predictions to demographic attributes. Experimental results show that baseline models achieve strong predictive performance, with accuracy ranging from 0.807 to 0.816; however, fairness evaluation reveals substantial disparities, with Disparate Impact as low as 0.190 and Demographic Parity Difference exceeding 0.27 in some cases. After applying the proposed multi-stage mitigation approach, fairness metrics improve significantly: Disparate Impact meets or exceeds the legal threshold of 0.80 across all models, with reductions in Demographic Parity Difference of 70–85% and Equal Opportunity Difference of 47–75%, and demographic disparities are reduced to 0.029–0.053. These improvements are achieved with minimal performance trade-off, as overall accuracy decreases by at most 4.2 % points while ROC AUC remains unchanged. The findings demonstrate that bias in recruitment systems arises from the interplay between data and model dynamics and highlight the importance of coordinated mitigation strategies throughout the machine learning lifecycle. This work provides a practical, scalable approach to developing fair and transparent AI systems for hiring applications.
Gideon Assafuah, Claude Turner, C. Turner et al.· International Journal of Adv...· 0 citations
Results indicate that global feature importance, used as an active search signal rather than a post-hoc diagnostic, improves both the effectiveness and the efficiency of individual fairness testing.
H. Mamman, Abdullateef Oluwagbemiga Balogun, Mustapha Maidawa et al.· Journal of King Saud Univers...· 0 citations
The integration of Artificial Intelligence (AI) into high-stakes social work, such as child welfare and criminal justice, promises efficiency but risks perpetuating systemic biases. This study addresses the critical trade-off between predictive accuracy and algorithmic fairness. Baseline evaluations of Extreme Gradient Boosting (XGBoost) models across three datasets (COMPAS, AFST, and a synthetic dataset) revealed strong predictive performance (AUC up to 0.89) but significant racial bias. As shown in the fairness metrics comparison, pre-mitigation Statistical Parity Difference (SPD) values indicated severe bias (COMPAS: –0.21, AFST: –0.15, Synthetic: –0.18). To resolve this, we propose a “Three-Stage Fairness-Aware Framework” integrating pre-processing, in-processing, and post-processing mitigation strategies. The application of this framework successfully reduced bias to ethical thresholds, with post-mitigation SPD values improving significantly to –0.07, –0.04 and –0.06, respectively, while incurring minimal accuracy loss (< 5.1% AUC reduction). Furthermore, the study validates the framework’s computational scalability and its alignment with strict regulatory standards like the EU AI Act and GDPR. These findings provide an evidence-based blueprint for equitable and legally compliant AI governance in public services.
Yih-Chang Chen, Chia-Ching Lin, Sedat Agan· Far East Journal of Electron...· 0 citations
The integration of artificial intelligence (AI) and machine learning (ML) into education has transformed how student performance is predicted and monitored. Despite these advances, concerns regarding fairness, transparency, interpretability, and potential demographic bias remain significant challenges in educational prediction systems. This study developed ethically aligned and interpretable ML models for predicting student academic performance using only behavioural engagement and academic context variables. The open-access xAPI-Edu-Data dataset containing 480 student records was obtained from Kaggle. Four supervised algorithms, Multinomial Logistic Regression, Decision Tree, Random Forest, and XGBoost, were implemented in Python 3.11 using Scikit-learn, SHAP, and LIME frameworks. To minimise data leakage, all preprocessing operations, including standardisation and categorical encoding, were embedded within a Scikit-learn Pipeline fitted exclusively on the training folds. A stratified train/validation/test split (70%/15%/15%) with random_state = 42 was employed, while hyperparameters were optimised using five-fold cross-validation on the training and validation sets only. Model performance was evaluated using accuracy, macro F1-score, ROC-AUC, Brier Score, and Expected Calibration Error (ECE). Results showed that Random Forest achieved the best overall performance and calibration, while Logistic Regression provided superior interpretability. Across all models, visited_resources and raisedhands consistently emerged as the strongest predictors of academic achievement, emphasising the importance of student engagement behaviours. The study demonstrates that fairness-by-design prediction systems that exclude demographic variables can support equitable and actionable early-warning interventions. The novelty of the study lies in its ethically grounded and deployment-oriented framework for interpretable educational prediction in resource-constrained contexts.
M. A. Ayanwale, I. J. Chikezie· Journal of Computer Adaptive...· 0 citations
Artificial intelligence (AI)-based prediction models, including risk scoring systems and decision support systems, are being increasingly adopted in health care. Addressing AI fairness is essential to fighting health disparities and ensuring equitable model performance and patient outcomes. However, numerous and conflicting definitions of fairness complicate this effort. In this Viewpoint, we aim to support the transition of AI fairness from theory to practice using appropriate fairness metrics. We assess the relation of 27 fairness definitions identified in the literature to the model's intended use, type of decision influenced, and ethical principles of distributive justice. Because of limitations in some notions of fairness, we argue that clinical utility, performance-based metrics (such as area under the receiver operating characteristic curve), calibration, and statistical parity are the most relevant group-based metrics for medical applications. Through two use cases, we show that different metrics might be applicable depending on the intended use and ethical framework. Our approach provides practical guidance for fair AI development, helping AI developers and assessors to evaluate model fairness and understand the effects of bias mitigation strategies, thereby supporting equitable AI-based implementations.
S. L. van der Meijden, Yuqing Wang, Madelena Y. Ng et al.· The Lancet Digital Health· 1 citation
Early-warning systems (EWS) in learning analytics are increasingly used to identify learners at risk of course withdrawal, but their deployment-critical properties are often under-reported once predicted scores are converted into intervention policies. This study develops a deployment-oriented evaluation protocol for course-withdrawal risk using the Open University Learning Analytics Dataset (OULAD). An early-window feature set was constructed from the first four weeks of learner activity and evaluated under a group-wise train–test split by course presentation. Multiple classifiers were benchmarked, including logistic regression, histogram-based gradient boosting (HGB), random forest, support vector machine, AdaBoost, K-nearest neighbors, XGBoost, LightGBM, and CatBoost. A calibrated HGB model was then retained as the main probabilistic model for downstream analyses of probability reliability, threshold sensitivity, subgroup fairness with bootstrap uncertainty, selective prediction, and capacity-based Top-x% alerting. Several tree-based and boosting models achieved comparable held-out discrimination, while calibrated HGB remained competitive across classification, ranking, and probability-reliability metrics. Threshold choices substantially changed the precision–recall balance, indicating that operating points should be treated as policy choices rather than universal defaults. Fairness audits showed policy-dependent observed group-level differences, especially in alert rates for disability status, although several subgroup error-rate and positive predictive value (PPV) differences remained uncertain. Selective prediction reduced risk on accepted cases as coverage decreased, whereas Top-x% alerting fixed outreach volume and made workload–effectiveness trade-offs explicit. Robustness analyses supported the 28-day window as a practical early-warning compromise and showed that absolute PPV values varied across held-out course-presentation splits. The findings suggest that EWS should be evaluated as policy-linked decision systems, integrating model benchmarking, calibration, fairness uncertainty, and capacity-aware decision rules before deployment.
Suhan Wu, Jingyi Duan, Min Luo· PLoS ONE· 1 citation