Fuzzy Classification Model Based on Genetic Algorithm

Olga Kochueva
The paper presents a new classification model based on a symbolic regression method, fuzzy inference system and genetic algorithm. For complex practical problems, building a unified predictive model for various states of a system or a process encounters a lot of difficulties, but the task can be divided into 2 stages: a)to obtain a classification of system states; b)to build models with good predictive qualities for each class. The fuzzy approach makes it possible to specify states (set of parameters) which can be assigned to more than one class. A feature of the presented model is the use of symbolic regression to identify variables that form the basis of the classification model and to clarify the interaction of parameters. The paper also presents an example of practical application of the model