ИНТЕЛЛЕКТУАЛЬНЫЙ АНАЛИЗ ПРОГРАММНЫХ АНОМАЛИЙ

The paper considers an actual applied problem related to the analysis of abnormal situations in program logs. An approach for recognizing anomalous events is proposed. The approach is based on constructing a program execution model in the form of a weighted digraph that displays the relationship between program functions and its transitions. A set of original algorithms for data preprocessing, analysis, and recognition has been developed. The algorithms effectively solve the stated problem.