PhishVQC: Optimizing Phishing URL Detection with Correlation Based Feature Selection and Variational Quantum Classifier

TL;DR

PhishVQC tackles the challenge of phishing URL detection by leveraging a quantum-classical hybrid framework that uses amplitude encoding and variational quantum classifiers with RealAmplitude and EfficientSU2 ansätze. The approach incorporates correlation-based feature selection to reduce inputs and encodes them into a quantum state, enabling learning with a parameterized circuit trained via COBYLA. On the PhiUSIIL phishing URL dataset, PhishVQC achieves macro F1 scores around 0.89 and demonstrates a 22% improvement over prior quantum studies, while highlighting significant wall-time costs as data scale grows. The work underscores the potential of quantum machine learning for cybersecurity and motivates future exploration of QSVM, QNN, and QCNN extensions as quantum hardware advances.

Abstract

Phishing URL detection is crucial in cybersecurity as malicious websites disguise themselves to steal sensitive infor mation. Traditional machine learning techniques struggle to per form well in complex real-world scenarios due to large datasets and intricate patterns. Motivated by quantum computing, this paper proposes using Variational Quantum Classifiers (VQC) to enhance phishing URL detection. We present PhishVQC, a quantum model that combines quantum feature maps and vari ational ansatzes such as RealAmplitude and EfficientSU2. The model is evaluated across two experimental setups with varying dataset sizes and feature map repetitions. PhishVQC achieves a maximum macro average F1-score of 0.89, showing a 22% improvement over prior studies. This highlights the potential of quantum machine learning to improve phishing detection accuracy. The study also notes computational challenges, with execution wall times increasing as dataset size grows.

Figures (5)

• Figure 1: Visual representation of the Bloch Sphere , the state of a single Qubit.
• Figure 2: The PhishVQC model for phishing detection.
• Figure 3: The RealAmplitude Ansatz.
• Figure 4: The EfficientSU2 Ansatz.
• Figure 5: Wall Time of Execution for PhishVQC with Different Setups.