Self-attention convolutional neural network optimized with season optimization algorithm Espoused Chronic Kidney Diseases Diagnosis in Big Data System

Abstract

Nowadays, Internet of Things (IoT) and cloud platforms are broadly employed in numerous healthcare applications. Instead of using the limited storage and processing power found in mobile devices, the vast amount of data generated by internet of things devices in healthcare sector can be evaluated on a cloud platform. In this article, the Self-Attention Convolutional Neural Network optimized with Season Optimization Algorithm is proposed for Chronic Kidney Disease Diagnosis using IoT and cloud computing in Smart Medical Big Data health care system (SACNN-SOA-CKD-IoT-CC). IoT devices, like wearable and sensors perform data acquisition process. For chronic kidney disease (CKD) diagnostic model, the self-Attention convolutional neural network (SACNN) is applied. But, the SACNN not divulge any optimization systems adoption to calculate the optimal parameters and to make sure the exact categorization of Chronic Kidney Disease (CKD). Therefore, the season optimization algorithm (SOA) is used to optimize SACNN. The proposed approach is implemented in python language its performance is analyzed with performance metrices, like sensitivity, accuracy, recall, f-measure, specificity, network latency, scalability, response time, delay, and accuracy. The proposed SACNN-SOA-CKD-IoT-CC method achieves higher accuracy in CKD dataset of 15.66 %, 21.65 % and 9.64 %, lower error rate of 11.27 %, 8.35 %, and 21.06 % compared to the existing methods, like intelligent internet of things with cloud centric clinical decision support scheme for the prediction of CKD (LR-AME-CKD-IoT-CC), diagnostic prediction method for CKD in internet of things (MLP-SVM-CKD-IoT-CC) and ensemble of deep learning based clinical decision support system for CKD detection in medical internet of things environment (EDL-CDSS-CKD-IoT-CC).