Тип публикации: статья из журнала

Год издания: 2018

Ключевые слова: fuzzy rule base, classification, differential evolution, large-scale optimization, unconstrained optimization

Аннотация: The paper presents an approach for fuzzy rule base design for classification problems, which consists in solving a single-objective unconstrained optimization task. To make this possible, the rule base is encoded by a set of sigmoid functions, 2 functions per variable, and every variable and rule has unique sigmoid functions. A speПоказать полностьюcial multicomponent variant of differential evolution is developed, with a particular type of mutation strategy assigned to each component. The algorithm is tested on a set of benchmark functions and classification problems, and shows competitive results compared to other approaches.

Журнал: International Journal on Information Technologies and Security

Выпуск журнала: Т. 10, № 2

Номера страниц: 79-88

ISSN журнала: 13138251

Место издания: Sofia

• Stanovov Vladimir (Reshetnev Siberian State University of Science and Technology)
• Akhmedova Shakhnaz (Reshetnev Siberian State University of Science and Technology)
• Semenkin Eugene (Reshetnev Siberian State University of Science and Technology)

• Ядро РИНЦ (eLIBRARY.RU)