COMPARATIVE ANALYSIS OF ALGORITHMS OF FUZZY PRODUCTION RULES’ DATABASE GENERATION WITH AN EXAMPLE OF SOLVING THE CLASSIFICATION PROBLEM

Soldatova Olga Petrovna, candidate of technical sciences, associate professor, sub-department of information systems and technologies, Samara State Aerospace University named after academician S. P. Korolev, op-soldatova@yandex.ru

004.032.26

The proposed method of estimation of consistency and completeness, based on the one used in the theory of graded formal logical systems the notions of syntactic and semantic conclusions. To optimize the structure of the resulting rule base allows selection of the type and settings of the parameters of membership functions and the elimination method overlapping and contradictory regulations. Studies were conducted on data from well-known problems of classification of iris flower data set obtained from the UCI repository.

fuzzy production rules’ database, membership function, algorithm Abe-Lan, rating algorithm, consistency and completeness, rules’ database optimization.