Software Identification by Standard Machine Learning Tools

This article considers tools for controlling software installed on personal computers of automated system users . The flaws of these software solutions are grounded, and an approach to identifying executable files with the help of a machine learning algorithm is developed and presented. This algorithm consists in the gradient decision tree boosting on the basis of such libraries as XGBoost, LightGBM, CatBoost. The identification of programs with the help of XGBoost and LightGBM is executed. The experimental results are compared with the results of earlier studies conducted by other authors. The findings show that the developed method allows for identifying violations in the adopted security policy during information processing in automated systems.