Predictive Threat Intelligence and Machine Learning for Risk Mitigation in Enterprise IT Security Migration Projects

• Abiodun Okunola

  Ladoke Akintola University Technology

  Author

Enterprise IT security migration projects face escalating cyber risks as organizations transition to cloud and hybrid infrastructures. Traditional reactive security measures, reliant on static rule sets and post-incident analysis, prove inadequate against sophisticated, rapidly evolving threats that emerge during migration windows. This study addresses the critical research gap in predictive threat intelligence (PTI) frameworks specifically designed for migration contexts, where attack surfaces expand exponentially and security controls undergo temporary reconfiguration. We propose a hybrid machine learning framework integrating LightGBM for threat classification with anomaly detection algorithms for real-time risk scoring. Using a comprehensive dataset comprising 5,420 migration project records and 780,000 security events from 120 enterprise environments (2021-2025), the framework achieved 91.2% accuracy in predicting high-risk migration events with an average lead time of 48 hours—significantly outperforming traditional signature-based detection (67.8% accuracy) and static risk assessment matrices (71.4% accuracy). The LightGBM classifier demonstrated superior performance with a 0.94 AUC and F1-score of 0.89, leveraging 47 behavioral features. These findings establish a replicable predictive framework for security practitioners, enabling proactive risk mitigation during critical migration phases. The study further contributes a validated methodology for integrating threat intelligence feeds into migration project governance, demonstrating that predictive analytics can reduce incident response times by 58% and breach-related costs by an estimated 42%. The framework's explainability through SHAP values enhances practitioner trust and adoption, bridging the gap between academic machine learning research and operational security practice.

• PDF

Copyright (c) 2026 Abiodun Okunola (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.