SINN Based Federated Learning Model for Intrusion Detection with Blockchain Technology in Digital Forensic
Priyanka Pramod Pawar, Deepak Kumar, Renata Krupa, Piyush Kumar Pareek, H M Manoj, K. Deepika · 2024
Today, smart devices permeate every part of our lives, ushering in a new technological era known as Internet of Things (IoT). There ration of cybersecurity risks due to the proliferation of smart devices in IoT contexts and the massive volumes of sensitive data they store. In order to determine the time and place of these attacks and to collect evidence that can lead to the identification of the perpetrators, digital forensics is essential. Research into digital forensics in an IoT setting is fraught with difficulties owing to factors such as the dispersed nature of data storage, the need to be able to track and verify the provenance of evidence, the difficulty in obtaining data from many sources, and the lack of visibility into the evidence collection procedure. This is why the research suggested merging two potential technologies to offer a complete answer. The research employed federated learning to train representations locally using data kept by IoT devices, with the help of a dataset that was created to mimic attacks on the IoT ecosystem. The study utilises the BoT-IoT dataset and employs a classification method called Shepard Interpolation Neural Network (SINN) with the Shuffled Frog-Leaping algorithm (SFLA). After that, in order to make the blockchain lightweight, the study aggregated data using blockchain by gathering parameters from the IoT gateway