Healthcare framework for identification of strokes based on versatile distributed computing and machine learning (jul 2023,10.1007/s00500-023-09002-1)

This paper presents a novel healthcare services framework that leverages versatile distributed computing and mobile cloud computing to enhance the delivery of medical care data. The framework is designed to integrate with diverse cloud settings and distributed computing resources, utilizing a Virtual-Dedicated-Server with 5 VCPU and 32 GB storage of RAM to develop a prototype specifically tailored for stroke patients with cardioembolic and cryptogenic subtypes, using Android-based mobile phones as the primary devices. The proposed framework consists of two fundamental application components: flexible usage and server request. By employing a Convolutional Neural Network (CNN) module, the system effectively distinguishes between the two-stroke subtypes. Simultaneously, the server application securely stores patient data, ensuring accessibility, security, and adaptability for stroke patients. The CNN module performs computations using the Stroke dataset, providing valuable insights into subtype classification. Additionally, the GBRF (Gradient boosting random forest) algorithm has been employed for further analysis and accurate predictive modeling within the healthcare system. The research contributions of this work are significant. Firstly, it introduces a unique healthcare mechanism that harnesses the potential of distributed computing and mobile cloud computing in improving healthcare services. Secondly, the paper verifies the exceptional value of stroke-related data collected through smartphones, underscoring the uniqueness and reliability of such data sources. Lastly, the system enables the provision of health status information to patients through a CNN application programming interface (API), facilitating seamless communication and personalized healthcare services. By integrating machine intelligence based technologies, the framework enhances the efficiency and effectiveness of stroke identification and contributing to the overall well-being of patients in a meaningful way.