Methods of forecasting and assessing the efficiency of information protection in complex data processing systems

The aim of the work is to develop a hybrid model for predicting vulnerabilities and assessing the effectiveness of information protection in DevOps-oriented systems. In the course of the work, a combination of machine learning methods (LSTM networks), statistical analysis (Monte Carlo method) and reliability theory (FTA, HARA) were used. Experimental studies were conducted on a distributed payment data processing system. As a result, it was possible to reduce the threat detection time by 34% and increase the accuracy of incident prediction to 89.2%. The protection efficiency coefficient increased from 0.67 to 0.83. The developed methodology for implementation in the CI/CD pipeline of a SaaS platform for the economic sector, which ensures its practical effectiveness.