Deterministic Boolean-Algebra Framework for Interpretable and Energy-Efficient Phishing URL Detection in Real Time

Phishing attacks present a critical cybersecurity threat, with global financial losses exceeding USD 70 million in 2024. Modern machine-learning-based detection methods achieve high accuracy but have fundamental limitations, including lack of interpretability, significant computational requirements, and high energy consumption, which restrict their use in resource-constrained environments. This research presents a novel deterministic approach based on Boolean algebra for detecting phishing URLs. The method employs a 15-dimensional Boolean feature space covering structural, protocol, content-based, infrastructure, and reputation-based features, formalized as mathematically rigorous logical rules. Experimental evaluation which based on a balanced dataset of 50,000 URLs demonstrated a classification accuracy of 89.1% along with substantial operational advantages: processing latency of approximately 1 ms (24–69× faster), power consumption of 4.8 mW (108–250× lower), and full decision interpretability unlike machine learning methods. The proposed Boolean approach enables transparent, energy-efficient, and high-performance threat detection suitable for real-time cybersecurity applications, establishing a foundation for next-generation security systems with verifiable detection mechanisms. The proposed system is not intended to replace advanced ML-based detection systems; rather, it serves as an additional first line of defense, providing rapid initial filtering with minimal resource overhead and forwarding complex or borderline cases to ML systems for secondary verification.