DETERMINING AN EFFECTIVE SET OF TIME SERIES DATA PARAMETERS SUITABLE FOR IDENTIFYING SECURITY INCIDENTS
Ескіз недоступний
Дата
2022
Назва журналу
Номер ISSN
Назва тому
Видавець
Bulgaria: University Publishing House “V. Aprilov”.
Анотація
The purpose of the research is developing a method of data analysis, systematization, and preparation of time series of sensors values for further analysis by the method of structural functions (Allan variance), which can be used to reveal characteristics of processes and their changes with high validity. The data collection was performed in a real cyberphysical system on large range time intervals using examples of 11 different temperature sensors. The analysis of the processed digital data was performed and influences factors on the way from the measurement point to the point of receiving information were determined. It is shown that the appropriate preparation of data arrays makes it possible to obtain distinctive generalized samples of time sequences parameters, the comparison to which allows to detect security incidents and abnormal processes in IoT sensor networks. The considering of these factors makes it possible to create a tool for sensors values pre-sorting and sample data sets obtaining. The developed hardware and software tools for such preliminary processing are described.
Опис
Ключові слова
time series data, data preparation, Identification of differences in date, Allan variance