Software Implementation of Takagi-Sugeno Fuzzy System Identification Methods Based on FCRM and Gath-Geva Clustering

Fuzzy logic algorithms are used to replace traditional control algorithms, when the processes of the control object are difficult to formalize and describe mathematically. The construction of fuzzy systems is done manually by experts, which is problematic in terms of resources/time, and automatically - here it’s important to test the software implementation. As a theoretical basis for this article, materials on clustering FCRM and Gat-Geva for identification of Takagi-Sugeno fuzzy systems were taken. The authors of the articles didn’t provide program code examples, which makes it impossible to test the described methods. In this article, the implementation of these clustering methods and the corresponding methods for Takagi-Sugeno fuzzy system identification is presented in the Python programming language. The resulting implementation was tested, and a comparative analysis of the accuracy of fuzzy systems built by implemented methods was carried out.