School of Information Technology and Engineering, Vellore Institute of Technology, Vellore 632014, India.
School of Creative Technologies, University of Bolton, Bolton BL3 5AB, UK.
Sensors (Basel). 2022 Jul 26;22(15):5574. doi: 10.3390/s22155574.
With the Internet of Things (IoT), mobile healthcare applications can now offer a variety of dimensionalities and online services. Disease Prediction Systems (DPS) increase the speed and accuracy of diagnosis, improving the quality of healthcare services. However, privacy is garnering an increasing amount of attention these days, especially concerning personal healthcare data, which are sensitive. There are a variety of prevailing privacy preservation techniques for disease prediction that are rendered. Nonetheless, there is a chance of medical users being affected by numerous disparate diseases. Therefore, it is vital to consider multi-label instances, which might decrease the accuracy. Thus, this paper proposes an efficient privacy-preserving (PP) scheme for patient healthcare data collected from IoT devices aimed at disease prediction in the modern Health Care System (HCS). The proposed system utilizes the Log of Round value-based Elliptic Curve Cryptography (LR-ECC) to enhance the security level during data transfer after the initial authentication phase. The authorized healthcare staff can securely download the patient data on the hospital side. Utilizing the Herding Genetic Algorithm-based Deep Learning Neural Network (EHGA-DLNN) can test these data with the trained system to predict the diseases. The experimental results demonstrate that the proposed approach improves prediction accuracy, privacy, and security compared to the existing methods.
随着物联网（IoT）的发展，移动医疗保健应用现在可以提供各种维度和在线服务。疾病预测系统（DPS）提高了诊断的速度和准确性，提高了医疗保健服务的质量。然而，隐私问题越来越受到关注，特别是涉及个人医疗保健数据，这些数据是敏感的。目前有多种流行的用于疾病预测的隐私保护技术。然而，医疗用户有可能受到多种不同疾病的影响。因此，考虑多标签实例是至关重要的，因为这可能会降低准确性。因此，本文提出了一种针对物联网设备采集的患者医疗保健数据的高效隐私保护（PP）方案，旨在为现代医疗保健系统（HCS）中的疾病预测提供支持。该系统在初始认证阶段后利用基于对数圆值的椭圆曲线密码学（LR-ECC）来提高数据传输过程中的安全性级别。授权的医疗保健人员可以在医院端安全地下载患者数据。利用基于羊群遗传算法的深度学习神经网络（EHGA-DLNN），可以利用训练有素的系统对这些数据进行测试，以预测疾病。实验结果表明，与现有方法相比，所提出的方法提高了预测准确性、隐私性和安全性。
