نبذة مختصرة : The article explores regression analysis methods as an effective tool for assessing factors that determine the stability of information security systems. In particular, it is shown that the use of multiple regression models allows us to quantitatively determine the impact of various technical, organizational and behavioral factors on the level of security of digital systems, predict the probability of cyber threats, assess the effectiveness of countermeasures and optimize the allocation of protection resources. Particular attention is paid to the application of the least squares method (LSM) to estimate the parameters of regression models, as well as the requirements that ensure the correctness of the obtained results, in particular, the independence of factor variables, the absence of multicollinearity, homoscedasticity, and uncorrelatedness of residuals. The article examines in detail the diagnostics of the main statistical violations: multicollinearity, heteroscedasticity and autocorrelation of residuals and methods for their elimination, including the transformation of variables, the introduction of auxiliary factors or the use of alternative estimation approaches, such as the generalized least squares method (GLS), the principal component analysis (PCA). It is shown that the systematic detection and correction of these violations are critically important for increasing the accuracy of forecasts and the reliability of conclusions in the field of cybersecurity, since erroneous assessments can lead to underestimation of risks or incorrect prioritization in threat management. The work demonstrates that the statistically sound application of regression analysis models creates a solid analytical basis for building intelligent decision support systems in the field of information security, increases the effectiveness of monitoring and contributes to the formation of strategic approaches to cyber risk management. The results obtained can be used both in scientific research aimed at improving the methodology for ...
Processing Request
No Comments.