Privacy aware-based federated learning framework for data sharing protection of internet of things devices

<p align="justify">Federated learning (FL) has emerged as one of the most effective solutions to deal with rapid utilization internet things (IoT) in big data markets. Through FL, local at each IoT device can be trained locally without sharing cloud server. However, this conventional FL may still suffer from privacy leakage when are trained, and model is shared server update global prediction model. This paper proposes a framework awareness protect including for devices. First, data/model encryption method using fully homomorphic introduced, aiming protecting privacy. Then, leveraging logistic regression approach discussed. Experimental results random datasets show that proposed obtain higher accuracy (up 4.84%) lower loss 66.4%) compared other baseline methods.</p>