Q-SSC³-IP: Quantum-Inspired Semi-Supervised Sparse Subspace Clustering with Contextual Coherence for Intrusion Prevention

Abstract The scale of large-scale, heterogeneous and high-dimensional network traffic has posed a serious challenge to the efficacy of conventional intrusion detection and prevention infrastructures, especially when the volume of labelled information is small, as well as when the attack patterns are constantly changing. Current supervised and deep learning-based security models have high labelling rates, lower generalization as well as higher false alarms in the face of stealthy or overlapping intrusion patterns. To overcome these shortcomings, this paper suggests Q-SSC³-IP as a quantum-inspired semi-supervised sparse subspace clustering scheme, with contextual coherence to intelligent network intrusion prevention. The suggested solution converts network traffic into a Hilbert feature space inspired by quantum computing, which allows the similarity modelling of interference to be performed to better distinguish between innocent and malicious connections. Semi-supervised sparse subspace learning mechanism is utilized to utilize labelled and unlabelled traffic samples in order to discover latent behavioural subspaces when under high-dimensional conditions. The regularization of contextual coherence also deals with temporal and statistical dependencies in between sequential connections to enable proactive prevention of intrusion other than passive detection. To prevent local minima, the quantum-inspired annealing is used in the optimization process in order to achieve stable convergence. The experimental analysis on NSL-KDD benchmark dataset show that the proposed framework yields a detection accuracy of 96.87, an F1-score of 96.11 along with a lower false alarm rate of 2.84, which is better than the state-of-the-art supervised, unsupervised and hybrid approaches to intrusion detection. The findings prove that Q-SSC³-IP is a scalable, explainable as well as robust solution to next-generation intrusion prevention systems that can be used under realistic network conditions