?
Методика анализа смыла текста с помощью семантической сети, содержащей нечеткие данные
The paper describes the developed methodology for fuzzy data extraction from natural language text and its analysis. The method includes the principle of the linguistic processor operation, the input data organization, the format of the generated data structure (semantic network integrating fuzzy data descriptions) and information search methodic. The analysis system is based on the object-attribute (OA) approach of the data and computing process organization. Fuzzy data are specified using a specially developed linguistic function (LF) formalism, which is an analogue of a linguistic variable. In the paper attention is drew to the implementation of LF by neural network, and the method of comparing output values of these LF during the information search in the semantic network.