An explainable AI-based intrusion detection system for cyber threats in stock markets using the SECURE framework

As electronic trading technology accelerates the way trades occur in financial markets, the ability to manipulate the markets in many different ways has be come stronger through this same technology. Two ways this happens are via spoofing and layering, which pervert the way prices are determined, confuse in vestors, and reduce the level of trust in the integrity of financial markets. Exist ing surveillance mechanisms were predominantly rule-based or relied on opaque machine learning models, limiting their adaptability and interpretability. This study proposed SECURE, an explainable artificial intelligence (AI)-based in trusion detection system designed to detect and interpret spoofing and layering activities in real time. The framework integrated ensemble machine learning models, cybersecurity-inspired monitoring logic, and shapley additive expla nations (SHAP)-based explainability to achieve accurate detection with trans parent decision-making. Using Level-II order book data, the system analysed behavioural and microstructural trading features to classify manipulative pat terns while generating feature-level explanations for each alert. Experimental evaluation demonstrated that the proposed framework achieved high detection accuracy, low false-positive rates, and dependable real-time performance. The results indicated that combining intrusion detection principles with explainable AI improved regulatory interpretability and strengthened the effectiveness of au tomated market surveillance systems.