A feature selection algorithm for PNN optimized by binary PSO and applied to smart city intrusion detection system

In the smart city construction and development process, network security is a vital link that must be addressed. As a critical technology to ensure the security of smart cities, intrusion detection systems play multiple roles, such as real-time monitoring, threat identification, and event response. By deploying and optimizing efficient intrusion detection mechanisms, smart cities can effectively resist various network attacks and ensure the safety and stability of urban operations. Therefore, this paper proposes a PNN feature selection model based on binary particle swarm algorithm optimization (BiPSO) to select features for network traffic data and remove a large amount of redundant information; then, a variety of classifiers (random forest, KNN, Adaboost, BPNN) are used to classify the data after feature selection. We use NSL-KDD for experiments. The experimental results show that the BiPSO-PNN proposed in this paper not only achieves accurate selection of data set features but also guarantees high accuracy on other classifiers, which has vital practical significance.