Automatic Threat Detection in ICT Systems by Selected Data Mining Methods and Software

The paper presents some real‑time analytical solutions that work in a proprietary‑designed system for IT security. It describes automatic methods of data transformations and analysis aiming at detection of potential threats (irregular system events, abnormal user behavior) both for time series and text documents without human supervision. Automation procedures used for time series and text documents are presented. Analyzed data was collected by Free Construction while protecting systems of electronic documents repositories (also including the Enterprise Content Management standards).