Abstract

This research presents a predictive framework for estimating the Length of Stay (LOS) of patients in clinical settings using machine learning integrated with Explainable AI (XAI). Traditional methods for estimating LOS rely heavily on clinical experience, which often results in subjective, slow, and occasionally inaccurate forecasts. By systematically analyzing multidimensional patient data, including demographic profiles, medical histories, pathological test results, and symptom severity, the proposed model delivers precise, data-driven predictions. A core contribution of this project is the application of Explainable AI, which illuminates the underlying decision-making process by quantifying the influence of specific features such as age, comorbidities, and diagnostic indicators. This transparency bridges the gap between complex algorithms and clinical intuition, fostering trust among healthcare professionals. Beyond individualized patient care, the system serves as a strategic tool for hospital administration, enabling optimized bed management, efficient staffing allocation, and streamlined treatment protocols. Ultimately, the framework supports a transition toward proactive hospital planning, reducing operational bottlenecks and enhancing the overall quality of patient care through intelligent, interpretable analytics.