Optimized BIGRU-Attention Network for OSA and Cardiovascular Disease Prediction with Risk Factor Classification

Cardiac illness is a rapidly rising worldwide killer. The development of an automated prediction framework for OSA-related cardiovascular problems is tough given that to the extensive information needed to effectively forecast heart illness. Several algorithms have been developed for use in IoT-based medical oversight systems, which utilize sensor data to identify ailments specifically.We deploy new learning models to improve prediction results. The existing methods lack the ability to forecast accurately, making them unsuitable for successful diagnosis. This study proposes BGEAN, a revolutionary framework that combines deep learning with computing devices to identify cardiac issues accurately. Data are pre-processed utilizing Studentized Residual Methods to address missing values and outliers. Improved Gated Recurrent Neural Networks help classify pre-processed data as normal or abnormal. Thorough testing and comparison to top deep learning techniques are conducted. Performance indicators, such as accuracy (99.99%), recall (99.35%), specificity (90.12%), precision (99.5), F1-score (99.99%), and Model Building Time (MBT) (0.212sec), are computed. The findings show that the technology that was suggested surpassed current algorithms in accuracy, error, and reaction time. With its many user situations, BGEAN provides the most diverse environment for forecasting cardiac diseases.