Deep Stacked Kronecker Network for Intrusion Detection System in IoT Environment with Big Data
J. Sreemathy, A. Arun, S. Sureshu, Shaik Mahaboob Basha · Cybernetics & Systems · 2025
The vulnerabilities present in Internet of Things (IoT) systems lead to security threats that can compromise smart environment applications. Thus, this research concentrates on the development of a Deep Stacked Kronecker network-based IDS called DSK-IDSNet for ensuring the security of big data in an IoT environment. In this case, nodes begin collecting data, which is then routed to the Base Station (BS) using the proposed Flamingo Cosine Optimization (FCO). The data received at the BS is then given to the Map Reduce module, data preprocessing and feature selection are carried out during the mapper phase, while intrusion detection occurs in the reducer phase. The preprocessing step selects the most relevant features using Motyka similarity. These features from the mapper modules are passed to the reducer, where the DSK-IDSNet, created by merging the Deep Kronecker Network (DKN) and Deep Stacked Autoencoder (DSA), is applied for intrusion detection. The devised model is validated using the BoT-IoT and the NSL–KDD database. The performance of the devised DSK-IDSNet is analyzed using the evaluation metrics namely, True Negative Rate (TNR), accuracy, True Positive Rate (TPR), F1 score, energy, and delay, which obtained values of 0.903, 0.931, 0.923, 0.799 J and 0.331 sec.