Enhancing COVID-19 Diagnosis Accuracy and Transparency with eXplainable Artificial Intelligence (XAI) Techniques

Abstract The COVID-19 pandemic has strained global healthcare systems, necessitating efficient screening tools for the timely identification and management of SARS-CoV-2 infections. This research proposes a novel approach that utilizes machine learning and XAI techniques to create an effective screening tool for COVID-19 [1]. By training a prediction model on a dataset of over 50,000 tested individuals, incorporating features like age, sex, contact history, and clinical symptoms, accurate estimation of SARS-CoV-2 infection risk is achieved. Integration of XAI techniques, such as feature importance analysis and decision boundary visualization, enhances transparency, improves accuracy, and fosters confidence among healthcare professionals [2-5]. The objective is to address comprehension challenges and effectively leverage diverse knowledge for diagnosis, empowering global medical staff with a valuable tool for triaging patients and optimizing resource allocation. This research significantly contributes to combating the COVID-19 pandemic and enhancing healthcare systems' resilience against similar challenges.