QDKFFHNet: Quantum Dilated Kronecker Feed Forward Harmonic Net for Intrusion Detection in IoT-Based Smart City Applications

Abstract

In today's world, people are worried about keeping their information safe and secure. One essential way to keep safe is by employing a technique called intrusion detection (ID), which helps to find threats in networks while data is being sent. Deep learning (DL) aids the Internet of Things (IoT) and smart cities by creating devices that can think and make decisions on their own without human help. Hence, an effective approach named Quantum Dilated Kronecker Feed Forward Harmonic Net (QDKFFHNet) is introduced for ID in IoT-Based Smart City Applications. Initially, the system model is considered, and then, data collection is carried out. Thereafter, data normalization is conducted by linear normalization. After that, feature dimension transformation is conducted by information gain. Feature extraction is conducted. Moreover, feature conversion is conducted, and feature selection is performed by Mahalanobis distance and Wave-Hedges metric. Lastly, ID is done by employing QDKFFHNet, which is the combination of Quantum Dilated Convolutional Neural Network (QDCNN) and Deep Kronecker Network (DKN). It is noticed that QDKFFHNet has gained an accuracy of 92.69%, a negative predictive value (NPV) of 86.73%, a specificity of 91.77%, a sensitivity of 92.79%, and a positive predictive value (PPV) of 92.60%.