Advanced security learning model for predicting network threats with intelligent approaches

Network intrusion detection (NID) has emerged as a research hotspot in recent years for several kinds of cyberspace security-related topics, which is a key component of national cyberspace security strategies. Creating fast and efficient intelligent NID methods that use cutting-edge machine learning (ML) algorithms is essential for defending against numerous types of NIs in complicated network environments. This research proposes the Enhanced Learning model for Network Threat Prediction (EL-NTP) where Density-based Clustering (DCN) with Noise and Kernel clustering (KC) are combined in this NI detection technique. Initially, hierarchical clustering is employed to handle network traffic data. Subsequently, to reduce the error accumulation during the development of EL-NTP and to improve the separation of the top nodes in the KC, a bottom-up merging method is used. Twin DCN and KC are then incorporated into the construct to utilize the NID model which effectively and top-down identifies the NI category. The efficacy of the proposed EL-NTP methodology in detecting intrusions is assessed utilizing the UNSW-NB15 and NSL-KDD benchmark datasets for ID. The experimental outcomes indicate that EL-NTP demonstrates comparable detection capabilities to several previously established NI detection (NID) methods and is proficient in recognizing a diverse array of NI types.