Assessment of the bank’s operational risk criticality level based on neural network technologies

Abstract

The article is devoted to the issues of controlling the operational risks of a credit institution arising in the process of using IT technologies. Among banking risks, operational risk occupies a special place, primarily due to the fact, that it affects various areas of banking activity and is difficult to separate from other types of risk. Operational risks arise, among other things, as a result of downtime or incorrect operation of technical systems and equipment. Due to the constant growth in the degree of automation of banking business processes, new IT risk groups are emerging that can have a significant impact on the activities of a credit institution. The aim of the work is to create an artificial neural network using the high-level Keras library in Python, which automatically controls the level of criticality of the IT risk that has arisen. In the article, based on the analysis of risk events associated with the use of IT technologies, the data flows entering the input of the neural network is identified and its structure is determined. The paper also presents the results of training a neural network created by the authors based on the generated data sets. The use of intelligent methods for assessing the level of criticality of operational IT risk allows you to quickly take measures to minimize the consequences, and thus reduce direct and indirect losses. In connection with the above, the automation of operational risk management based on the use of neural network technologies is currently one of the most urgent tasks for credit institutions. The results obtained are new and can be used by credit institutions in the process of building automated systems for monitoring and managing operational risks.