An Intelligent Intrusion Detection for Smart Cities Application Based on Random Optimization with Recurrent Network

Regardless of their power, unsecured smart society networks have turned out to be potential back door entry points. An Artificial Intelligence (AI) detection technique heavily relies on the nature of the input features. The supervised learning approach has the detection strategy to find the connection between the feature and its interruption level. In this chapter, a smart city application-based Intrusion Detection (ID) system is assessed using a deep learning model that uses a Recurrent Neural Network (RNN) with Random Monarch Butterfly (RMB) optimization techniques to take complex state features from raw data in an unaided manner and handle genuine data representation produced from smart meters and sensors using previously mentioned optimization. The chosen features of KDD cup dataset in the RNN with Gated Recurrent Unit (GRU) was used to classify the testing data as interrupted or not if interruption indicates we detect the attacks. Finally, the data are stored in the cloud server to upgrade the security level. The experimental assessment results showed the proposed procedure outperformed conventional techniques.