A methodology for the identification of extremal loading in data flows in information systemsстатья
Информация о цитировании статьи получена из
Scopus
Статья опубликована в журнале из списка Web of Science и/или Scopus
Дата последнего поиска статьи во внешних источниках: 8 февраля 2017 г.
Аннотация:The paper presents two techniques for the identification of extremal loading via determination of special thresholds in order to distinguish between "normal" and "extreme" values in information data flows. Both algorithms are based on the Renyi limit theorem on rarefaction of renewal processes flows and the
Pickands-Balkema-de Haan theorem on the asymptotic distribution for peaks over large thresholds. The methodology can be applied to various information systems. The two methods differ by the direction of threshold moving. The ascending algorithm increases the value of the threshold, while the descending one decreases it step-by-step. In addition, for the descending method we suggest a way to process the cumulative data. The key stages of both methods are represented by the flowcharts. Some graphical results are demonstrated for test data generated by a special information system.