FIELD: computer equipment.
SUBSTANCE: method of classifying traffic, according to which a natural language message is received, determining a message scheme, dividing the message into meaningful syntactic and semantic fields, providing a plurality of nodes for analysing significant message fields, to which text analysis rules are applied depending on the application message protocol, wherein to classify a message by certain categories, an informal row-string of the informal search language and substring processing in a row is determined, and a probabilistic coefficient of relating the message to a specific category, converting an informal string pattern into a regular expression of a standard search language and processing a substring in a row, converting standard language regular expression to text analysis rules configuration module containing data and string processing methods, dynamically creating a plurality of program modules for configuring analysis rules and distributing them over a plurality of nodes for analysing significant message fields, applying analysis rule configuration to significant message fields, sending a message to the input of the neural network.
EFFECT: technical result consists in improvement of flexibility and reliability of messages classification.
8 cl, 7 dwg
