Extracting Significant Information from

Natural-Language Text with combining it in form of user pattern.

Igor P. Kuznetsov I.P., Matskevich A.G.

Moscow, Institute for Informatics Problems of the Russian Academy of Sciences




A system for extracting significant information from natural language text in Russian is considered. The system uses the methods which is based on the fact that the usual texts abounding synonymous ways of expressing the same content. This kind of system should be provided with the features for recognizing word combinations of same meaning, for object identification, for analysis of word role in object description and so on.

The paper describes an approach to a linguistic processor providing morphological, syntactical and context analyses on the level of semantic networks. The further processing is performed by logical-analytical means and methods which are based on the conception of semantic filters. They use the SUB-tree and semantic space of words for searching the significant information from free text