@article{4733,
  author = {Martin Lopez Nores},
  title = {Statistical and Predictive Evaluation of Cybersecurity Threat Dynamics and System Response Latency},
  journal = {Journal of Information Security Research},
  year = {2026},
  volume = {17},
  number = {2},
  doi = {https://doi.org/10.6025/jisr/2026/17/2/57-75},
  url = {https://www.dline.info/jisr/fulltext/v17n2/jisrv17n2_1.pdf},
  abstract = {The rapid evolution of cyber threats has exposed the limitations of traditional, reactive cybersecurity
frameworks, necessitating data driven and adaptive defense mechanisms. This study investigates the dynamics
between threat characteristics and system response latency through a comprehensive statistical and
predictive analysis of 1,000 cybersecurity incident records. Employing chi-square tests, correlation analyses,
effect size measurements, and Ridge regression modeling, the research evaluates how variables such as
threat severity, access level, and detection status influence breach occurrence and response time. Results
reveal that threat severity and privilege levels exhibit negligible associations with breach likelihood and
response latency, challenging conventional risk prioritization assumptions. Instead, threat detection status
emerges as the dominant predictor of response time (n²  0.76), indicating that current systems operate
primarily on reactive, event triggered paradigms rather than proactive, severity aware mechanisms.
Furthermore, the analysis identifies critical data quality challenges, including zero inflated response time
distributions and timestamp induced feature leakage, which artificially inflate predictive metrics and
compromise model generalizability. These findings underscore a significant gap between theoretical
cybersecurity frameworks and practical system behavior. The study concludes that enhancing threatdetection
accuracy, implementing rigorous data-preprocessing protocols, and integrating meaningful
temporal and contextual features are essential for developing robust, proactive cybersecurity architectures.
Future research should prioritize real world dataset validation and advanced machine learning techniques
to overcome current analytical constraints.},
}