As fraud and security breaches are becoming more frequent and sophisticated, traditional security solutions are not able to protect company assets. The problems which musts be solved to provide high security level for data and users of enterprise computer systems is related to analysis and correlation of  large amounts of real-time data from network and security devices. The high level context of data is often crucial element on the way of successful management of internal and external security threats or improvement of incident response time and compliance reporting. The network monitoring can be performed by various types of intrusion Detection systems which monitor and analyze network traffic to detect, identify, and report on suspicious activity or intrusions. But new security threats can be recognized only when we provide efficient way for pattern/anomaly recognition on larger volumes and greater variety of data. Another critical issue for each enterprise is to detect and prevent fraudulent activities by internal or external parties. This is why for security application we need new methods and tools to Improve risk assessment and associated scoring by building sophisticated machine learning models that can take into account hundreds or even thousands of indicators and complex patterns introduced by cooperation of systems and people and detectable only on using higher levels of network dependencies.