Adaptive Neural Network Filtering for Operational Assessment of Critical Resource Security

The object of the study is a new methodological and practical approach to solving the problem of adaptive neural network filtering. A hybrid adaptive approach to the operational assessment of the security of critical resources is described, which combines traditional Kalman filtering methods with the capabilities of artificial neural networks with training. An analysis of this approach features is made, it allows training and adjusting the filtering weights to the statistical characteristics of the security indicators of critical resources, measured and observed both linearly and non-linearly.