ASSESSMENT OF STABILITY FACTORS OF INFORMATION SECURITY SYSTEMS USING REGRESSION ANALYSIS METHODS

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