Optimizing intrusion detection using intelligent feature selection with machine learning model

Network security is a critical aspect of information technology, targeting to safeguard the confidentiality, integrity, and availability of data transmitted across computer networks. Intrusion Detection Systems (IDS) plays an essential role, serving as vigilant sentinels against illegal access, malicious actions, and potential threats. IDS operates on analysing the network or system activities, analyzing patterns, and detecting anomalies that may specify security breaches. The enhancement of network security via the integration of feature selection and machine learning, particularly in the context of IDS. Feature selection methods enable the identification and prioritization of key data attributes, optimizing the performance of machine learning algorithms by focusing on relevant information. Machine learning algorithms, such as decision trees, support vector machines, or neural networks, leverage the chosen features to dynamically adapt and learn from evolving cyber threats. Therefore, this study develops a new gravitational search algorithm-based feature selection with optimal quantum neural network (GSAFS-OQNN) model for intrusion detection and classification. The proposed GSAFS-OQNN approach lies in the effectual detection of intrusions. To accomplish this, the GSAFS-OQNN method exploits a Z-score normalization approach at the preprocessing step. Furthermore, GSAFS-OQNN technique designs the GSAFS model to derive an optimum subset of features. For intrusion detection, quantum neural network (QNN) is applied. Finally, the sandpiper optimization (SPO) technique is used to finetune the parameters of the QNN model. The experimental analysis of GSAFS-OQNN model is implemented on benchmark IDS datasets. The comprehensive results stated the betterment of GSAFS-OQNN method over recent approaches.