Optimized Smart City Communication Security using Pacenids Classification Technique with WB-Smote Normalization

Ensuring the security of communication networks is crucial for preserving urban resilience in the fast-changing smart city environment. This study presents an Enhanced method of Parallel ABILSTM and CBIGRU Ensemble Network Intrusion Detection System (EPACENIDS) for network intrusion detection. The system has been carefully built to deal with many kinds of intrusions, such as Denial of Service(DoS), Probe, Remote to Local (R2L), and User to Root (U2R), which threaten the communication networks of smart cities. Furthermore, preprocessing methods like Z-Score, Principal Component Analysis (PCA), and Weighted Borderline-Synthetic Minority Over-sampling Technique(WB-SMOTE) have been integrated into the comparison framework to assess their influence on overall performance. Important parameters including recall, accuracy, precision, False Alarm Rate (FAR), specificity, and F-measure have been used to evaluate EPACENIDS performance. The suggested approach is compared with conventional techniques like Long-Short Term Memory (LSTM), and Bi Directional LSTM to provide an accurate comparison. According to the experimental results, EPACENIDS performs better than the other approaches especially when combined with SMOTE preprocessing on the NSL-KDD dataset. The results highlight the effectiveness of SMOTE-PCA-PACENIDS in enhancing the system's overall performance and highlight the resilience of EPACENIDS in smart city environments.