At considerable expansion of information streams generally containing not structured data, there is a problem of automated allocation of information fragments not corresponding to the general context of a data flow.
One of possible methodologies suitable for the given problem is the content-analysis, and the semantic filters constructed on its basis.
At the moment the systems investigating information streams of the Internet network are information retrieval systems.
In functioning algorithms of these systems methods of semantic networks, content-analysis and Text Mining are used .
For the effective decision of a problem of allocation of fragments it is necessary to design a system based on the use of high-intellectual schemes of the analysis of data semantics (definition of fragments which in some sense differ from a context