FIELD: computer equipment.
SUBSTANCE: invention relates to computing equipment for natural language processing. Technical result is achieved by identifying with the first model of the classifier for processing the first set of classification attributes extracted from semantic-syntactic structures, the set of root components, such that each root component from the set of root components is associated with a span from a set of spans, where a span is a fragment of a text, and each span represents an attribute of an information object of a certain ontological class; identification using the second classifier model for processing the second set of classification attributes extracted from the semantic-syntactic structures, child components of each component from the set of root components; and determining, using the third classifier model for processing the third set of classification attributes extracted from the semantic-syntactic structures, whether the first span of the multiple spans and the second span of the multiple spans are associated with the same information object.
EFFECT: technical result is the higher efficiency of natural language processing in terms of identifying information objects and the relations between them.
20 cl, 19 dwg
