Preserving Privacy of IoT Healthcare Data using Differential Privacy and LSTM

Abstract

The Internet of Things (IoT) is a powerful technology creating revolutions in multiple industries for ex: Traffic and Healthcare domains. The patient data collected by continuous monitoring using IoT will support in treating the patients and make a positive impact on patients' well-being and increase the efficiency of healthcare workers. It is crucial to be aware of certain drawbacks and risks associated with protecting the privacy of the patient data which is one of the major problems being faced in the healthcare domain. Harmful individuals/Agencies will use IoT devices to obtain private data of patients. It’s of prime importance to protect privacy in healthcare. To improve the privacy of IoT Healthcare data, Geometric data perturbation along with Noise addition is introduced in this study utilizing Laplace Noise which comes under the framework of Differential Privacy. To increase accuracy, a deep learning technique Long Short-Term Memory (LSTM) is applied in this paper. LSTM has proven to be a superior model in accuracy when compared with other models like Decision Tree and Naive Bayes.