Team Work Optimizer Based Bidirectional LSTM Model for Designing a Secure Cybersecurity Model

The Internet's rapid growth and the volume of data being transmitted over it have been accompanied by a steady increase in threats to network security. Hackers attempt to steal confidential information or stop vital computer functions by targeting vulnerabilities in network infrastructure. Network Intrusion Detection Systems (NIDS) aim to quickly detect and prevent intrusions on network resources. NIDS uses a number of supervised and unsupervised machine learning techniques to address network security issues and spot anomalies in network traffic. One of the datasets used to train these NIDSs is attack traces. Sadly, these defenses are unable to keep up with how modern attacks are changing. For this reason, training and developing NIDS requires access to a cutting-edge, comprehensive dataset that contains both common and attack actions from the present day. This study uses the Bidirectional Long-Short Term Memory (BLSTM) model to predict different types of attacks. The BLSTM weight is optimally selected by the Team Work Optimisation Algorithm (TWOA) in order to improve classification accuracy. Experimental testing on a single benchmark network intrusion dataset shows that the proposed method performs better than other popular ML and DL models as well as state-of-the-art approaches. In particular, the proposed method achieved a maximum accuracy of 98.79% in detecting network assaults when tested on the SDN-IoT dataset.