Method of protection of personal data in its processing in information system based on the artificial neural network

Abstract

One of the most actively developing areas of information security is the User Behavior Analytics. This paper presents a method of detecting anomalies in the behavior of an information system user has been developed, based on the use of an artificial neural network that signals the commission of illegal actions. Users behavior characteristics had been offered to use sample input values: access time; duration of work performed; place of access; a set of data with which the user works; list of actions taken. An approach to assigning numeric values to user characteristics is proposed, based on the fuzzy set theory and One-Hot Encoding method. Method provides more effective detecting abnormalities in user behavior than analyze by information security specialist without using the special automation tools.