Abstract

Decision-support systems (DSS) play a critical role in domains such as healthcare, finance, and industrial automation, where data-driven recommendations guide human decision-making. While complex computational models, including deep learning and ensemble methods, provide high predictive accuracy, their black-box nature often hinders trust, accountability, and compliance with regulatory requirements. This paper investigates the development and application of interpretable computational models to enhance transparency in decision-support systems. Techniques such as rule-based models, decision trees, generalized additive models, attention mechanisms, and explainable AI frameworks are examined for their ability to balance predictive performance with interpretability. The study focuses on methods for extracting feature importance, generating model explanations, and designing human-understandable representations of complex models. Experimental evaluations demonstrate that interpretable models can achieve comparable performance to black-box models while providing clear insights into decision rationale. Furthermore, the research explores challenges in integrating interpretability into real-time decision-making environments and addresses trade-offs between transparency, complexity, and predictive accuracy. By promoting model transparency and trust, interpretable computational models facilitate accountable decision-making, improved user adoption, and enhanced alignment with ethical and legal standards