The widespread adoption of Bring Your Own Device (BYOD) policies in blue-collar industries introduces cybersecurity risks that differ substantially from white-collar environments. Remote workers in logistics, transportation, and manufacturing often lack the technical literacy to recognize and respond to security threats on personal mobile devices, yet existing frameworks have been designed primarily for resource-rich, white-collar contexts. This paper proposes a hybrid BYOD security risk mitigation framework for remote Android users in blue-collar enterprises, integrating three components: a CHI-IG hybrid feature selection pipeline that reduces 57 raw features to 39 selected features (31.6% reduction); the Modified Classification and Regression (MCAR) algorithm, a confidence-weighted ensemble combining Random Forest and Gradient Boosting evaluated under 10-fold stratified cross-validation against four baseline classifiers; and a contextual just-in-time user education module delivered at the point of threat detection. MCAR achieves 98.13% accuracy, precision, recall, and F1-score (95% CI: ±0.40%). A full 17-configuration ablation study confirms the independent contribution of CHI-IG preprocessing and demonstrates that MCAR uniquely supports a threshold-based human escalation mechanism through calibrated probability outputs. The framework addresses a gap identified in systematic literature reviews and provides an integrated, computationally efficient solution for low-resource, low-literacy BYOD deployment environments.
Alqahtani et al. (Fri,) studied this question.