Porównanie
wybranych metod statystycznych i metod sztucznej inteligencji do przewidywania
zdarzeń w oprogramowaniu zabezpieczającym systemy przechowywania
dokumentów cyfrowych, w tym systemy klasy Enterprise
Content Management
A comparison of some statistical methods and artificial
intelligence methods for predicting events in software protecting digital
documents repositories, including Enterprise Content Management
Author(s): Kamil Sapała, Marcin Piołun-Noyszewski, Marcin WeissSubject(s): Economy
Published by: Wydawnictwo Uniwersytetu Ekonomicznego we Wrocławiu
Keywords: neural networks; ARIMA; exponential smoothing; real-time analysis
Summary/Abstract: Recently statistical analysis of IT security events has been focusing more attention. Analytical modules have more often been implemented in systems protecting companies from security threats. In this field automation and analysis without human supervision are of great importance. The paper presents a performance of automatic expert modules applied to predict time series, if its quantities were unknown. Created models without appropriate time series modification procedures and correct specification of parameters work only in a limited way. Nevertheless, the predictions of seasonal time series can provide valuable information about potential security threats to a company.
Journal: Prace Naukowe Uniwersytetu Ekonomicznego we Wrocławiu
- Issue Year: 2017
- Issue No: 469
- Page Range: 159-166
- Page Count: 8
- Language: Polish