Среда имитационного моделирования генерации лог-файлов при мониторинге событий

Abstract

Aim. The aim of the work is to solve the problem of modeling the generation of data sets in monitoring systems. Automatic detection of abnormal events in processes in real time is an urgent and important task. Materials and methods. To identify various kinds of anomalies, various data mining methods are often used. Often, to obtain a high accuracy of the analysis, preliminary training is required using a large amount of data and representative samples. However, it is not always possible to obtain a representative sample of various kinds of anomalies from real sources. The developed environment makes it possible to consider various scenarios of movement of given objects of observation and generate data containing various parameters that can be provided in such systems. Time indicators are considered when an object moves from one monitoring point to another, time delays when registering one event by several monitoring devices. The ability to add anomalous events to scripts has been developed. Results. Experiments have been carried out to simulate data generation in the implementation of motion scenarios for objects of observation. The proposed environment is intended for use by researchers in the field of data processing and analysis for the development and testing of new methods of data processing in monitoring systems. Conclusion. The data sets obtained during the experiments can be used as reference data for training data analysis mechanisms, and for testing various analysis mechanisms in order to assess the accuracy of identifying possible deviations.