METHOD FOR IDENTIFYING OBJECTS CORRESPONDING TO THEIR DESCRIPTIONS

摘要

computing; data processing. SUBSTANCE: method involves applying linguistic sorting of all words containing in the given text in predefined clusters for building its thesaurus. Statistical processing of the words is carried out in each cluster of the given text. Linguistic sorting of all words belonging to the sample from prepared query similar to the linguistic sorting done when creating text thesaurus. Statistical thesaurus coincidence measure is calculated on the words sampled from prepared query and texts retrieved in a selected database. Decisions concerning some object identification are drawn by comparing the calculated statistical coincidence measures built for different texts. EFFECT: wide range of identification features; usability of unrestricted volumes of additional data. 11 cl