Abstract:
Linguistic processor allocating information objects and relations from texts of a natural language (NL) with formation of knowledge structures is considered. Experience of construction of such processors in various subject domains is analyzed. Owing to a high variety of forms which are used in NL for description of various objects and phenomena, construction of whatever complete «language model» is an insoluble problem. Therefore an important component of the linguistic processor which is carrying out formalization of texts stream is adjustment to features of entrance texts. In article such means are examined at level of syntactic and semantic analysis.