A MODEL OF STRATEGY ANALYSIS DURING THE DYNAMIC INTERACTION OF PHISHING ATTACK PARTICIPANTS

Abstract

The paper proposes an approach that allows countering attacks on cryptocurrency exchanges and their clients. This approach is formalized in the form of a synthesis of a dynamic model of resistance to phishing attacks and a perceptron model in the form of the simplest artificial neural network. The dynamics of the confrontation are determined by a system of differential equations that determines the change in the states of the victim of phishing attacks and the attacker who organizes such attacks. This allows to find optimal strategies for opposing parties within the scheme of a bilinear differential game with complete information. The solution of the game allows you to determine payment matrices, which are elements of the training set for artificial neural networks. The synthesis of such models will make it possible to find a strategy to resist phishing with a sufficient degree of accuracy. This will minimize the losses of the victim of phishing attacks and of the protection side, which provides a secure system of communication with clients of the cryptocurrency exchange. The proposed neuro-game approach makes it possible to effectively forecast the process of countering phishing in the context of costs for parties using different strategies.