Оптимізація параметрів моделі прогнозування міцності бетону шляхом навчання та тестування

Abstract

Predicted concrete strength’s model optimization based on fuzzy logic, with model training has been offered in the “MATLAB” environment. Model training was performed by adjusting the concentration ratios of components’ membership functions and weighting rules of each knowledge base. Obtained discrepancy values of training data set show sufficient sensitivity of concrete strength prediction model to study.