The Fragility of Phishing Detection Models: Evidence from Cross-Corpus Transfer, Prevalence Shift, Artifact Learning, and Evasion Risk
Open Access
- 29 June 2026
- journal article
- research article
- Published by MDPI AG in Big Data and Cognitive Computing
- Vol. 10 (7) , 211
Abstract
Phishing detection models often report strong benchmark performance, yet their reliability under realistic deployment conditions remains uncertain. This study examines this problem by investigating three failure modes of cross-dataset phishing email detection: corpus generalization failure, asymmetric prevalence-shift failure, and artifact-driven spurious learning. Using six public email corpora, CEAS_08, Enron, Ling, Nazario, Nigerian Fraud, and SpamAssassin, the study evaluates Term Frequency (TF) and Inverse Document Frequency (IDF)-based Logistic Regression and Linear Support Vector Classifier (SVC) models across pooled baseline testing, single-corpus cross-dataset transfer, leave-one-corpus-out pooled training, prevalence-shift simulation, training prevalence manipulation, dataset-identification analysis, top-feature inspection, artifact-removal ablation, and targeted feature-sensitivity masking. The findings show that single-corpus models are unstable under cross-dataset transfer, with F1-scores varying substantially across source–target combinations. In contrast, leave-one-corpus-out pooled training improves robustness, with Logistic Regression achieving sustained F1-scores between 0.8201 and 0.8994, and Linear SVC achieving F1-scores between 0.7607 and 0.8910 across unseen corpora. Prevalence-shift experiments reveal that failure is asymmetric and threshold-dependent. High-prevalence-trained models maintain high recall under fixed thresholds but suffer sharp recall degradation when operational alert-budget constraints are imposed. Conversely, low-prevalence-trained models become overly conservative in high-threat environments, producing high precision but substantially lower recall and poorer calibration. Artifact analyses further show that source corpus identity is highly learnable, with dataset-identification accuracy reaching 0.9722 for Logistic Regression and 0.9806 for Linear SVC. Top-feature and masking analyses indicate that models rely partly on corpus markers, date tokens, URL/domain terms, headers, and other artifact-like features rather than only general phishing indicators. The study contributes a deployment-aware and adversary-aware evaluation framework for phishing detection. It shows that benchmark accuracy alone is insufficient for assessing real-world robustness and that reliable phishing detection requires cross-corpus validation, prevalence-aware thresholding, and systematic testing for artifact-driven spurious learning.Keywords
All Related Versions
This publication has 30 references indexed in Scilit:
- Evaluation of performance measures in predictive artificial intelligence models to support medical decisions: overview and guidanceThe Lancet Digital Health, 2025
- Business email compromise: A systematic review of understanding, detection, and challengesComputers & Security, 2025
- Artificial intelligence and machine learning in cybersecurity: a deep dive into state-of-the-art techniques and future paradigmsKnowledge and Information Systems, 2025
- The Harms of Class Imbalance Corrections for Machine Learning Based Prediction Models: A Simulation StudyStatistics in Medicine, 2025
- Staying ahead of phishers: a review of recent advances and emerging methodologies in phishing detectionArtificial Intelligence Review, 2024
- Novel interpretable and robust web-based AI platform for phishing email detectionComputers and Electrical Engineering, 2024
- PhishingRTDS: A real-time detection system for phishing attacks using a Deep Learning modelComputers & Security, 2024
- Classifier calibration: a survey on how to assess and improve predicted class probabilitiesMachine Learning, 2023
- A systematic literature review on phishing website detection techniquesJournal of King Saud University Computer and Information Sciences, 2023
- Applications of deep learning for phishing detection: a systematic literature reviewKnowledge and Information Systems, 2022