Skip to content
Open access

Explainable Artificial Intelligence and Machine Learning for Responsible Decision-Making: A Comparative Study of Model Accuracy, Transparency, and Fairness

Nov 2026 · International Journal of Artificial Intelligence, Machine Learning and Data Analytics · 0 citations · 18 references

Abstract

Explainable Artificial Intelligence (XAI) has emerged as an important research area in developing transparent, accountable, and trustworthy machine learning systems. Although machine learning models can achieve high predictive accuracy, their complex decision-making processes may limit interpretability and create challenges concerning fairness, accountability, and responsible implementation. This study aims to comparatively evaluate artificial intelligence and machine learning models based on three critical dimensions: predictive accuracy, transparency, and fairness. The proposed research examines interpretable algorithms, including Logistic Regression and Decision Trees, alongside complex models such as Random Forest, Gradient Boosting, and other ensemble learning techniques. Explainability methods, particularly SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME), are considered for interpreting model predictions and identifying influential features. Model performance will be assessed using accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve, while transparency and fairness will be evaluated through explanation quality, interpretability, and appropriate group fairness metrics. The study also investigates potential trade-offs between predictive performance, model complexity, and equitable outcomes. Drawing on established research in explainable AI, explanation evaluation, and algorithmic fairness, the research proposes a comparative framework for identifying models suitable for responsible decision-making. The findings are expected to support researchers, developers, and organizations in selecting machine learning approaches that balance predictive effectiveness with transparency, fairness, and accountability.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.