Methods of Assessing the Effectiveness of Network Content Processing Systems for Detecting Malicious Information, Taking into Account the Elimination of Uncertainty in the Semantic Content of Information Objects

The paper presents a mathematical model of the process of changing the state of indicators (features) of harmful information to he detected as part of the functioning of network content processing systems. The model is based on Markov chains and takes into account the probability-time mechanism for changing the state of the functioning process of systems of this class in dynamics. On the basis of the model, a technique of multi-criteria dynamic evaluation of the effectiveness of network content processing systems is proposed. It is based on the analysis of current deviations of the indicator values of harmful information in dependence on the requirements to them. Also it uses and eliminates the uncertainty of the semantic content of information objects. The methods (technique) makes it possible to take into account transient processes by introducing controls of the structure, parameters and modes of functioning of network content processing systems under conditions of various types of influences.