Sarcouncil Journal of Engineering and Computer Sciences

Abstract: The integration of Explainable AI into health data analytics of big data could guarantee the detection of heart disease in intricate EHRs, thereby increasing the transparency of the models, clinical decision-making, and individualized care in terms of interpretable predictions and state-of-the-art feature selection methods. This is particularly relevant for wearable ECG devices. Using data from wearable devices that detect the electrocardiogram (ECG) in real-time, the authors of this study provide an explainable AI method for predicting cardiac illness using high-dimensional electronic health records (EHRs). Data cleansing, data normalization, feature selection, and dimensionality reduction are some of the exacting procedures that make up the preprocessing of EHR data in the suggested architecture. Reduces computation overhead and increases efficiency through the use of advanced preprocessing techniques such as Recursive Feature Elimination (RFE) for feature selection and Principal Component Analysis (PCA) for dimensionality reduction. To make accurate and scalable disease predictions using a Recurrent Neural Network (RNN) model, it is crucial to extract time-based patterns from electrocardiogram (ECG) samples. The model's 92% accuracy and strong performance on all robustness indicators precision, recall, and F1-score make it an attractive option. To achieve the goals of explainable AI in healthcare analytics, the suggested methodology would facilitate the development of trustworthy clinical decision support systems by promoting interpretability and transparency.