Accelerometer-derived physical activity, sarcopenia, and grip strength as modifiers of type 2 diabetes risk.

The Internet of Things (IoT) signifies an important evolutionary period in technology that affects nearly every sector, including homes, transportation, cities, athletics, commerce, education, or industry. An IoT comprises of interconnected devices enabling users and organizations to be more productive, efficient and allow time, energy, and financial savings. By 2030, the total number of IoT devices is expected to reach nearly 50 billion, data generated from these devices is expected to reach 79.4 zettabytes by 2025, and worldwide revenue from IoT is expected to be $3 trillion in 2026. Although more convenient, IoT also presents severe challenges, especially with security and privacy. Oftentimes IoT devices are limited in resources and do not offer security protections resulting in easily compromised devices. As IoT uses a centralized server for consolidating and storing data, using a cloud-based service raises security concerns regarding attack targets. Additionally, IoT systems often utilize machine learning (ML) and deep learning (DL) algorithms as part of their system to produce real-time decisions, which may also heighten vulnerabilities as these models face threats from multiple attacks. In order to tackle privacy issues, primarily in the healthcare context, this article proposes a cluster-based data anonymization algorithm, which is conjoined with Federated Learning (FL). This is important as sensitive healthcare data is derived from IoT devices, making this a concern with the possibility of privacy being breached. The solution prioritizes IoT applications while harnessing the value of data analytics. As a result, this outlines an innovative approach to mitigating the risks involved with IoT adoption.