A study on the application of regression trees and adaptive neuro-fuzzy inference system in glass manufacturing process for packaging

Abstract

ANFIS and other algorithms were used for the classification of the defects that occur in the production process of glass for packing. In this study we used the Project “Newglass” installed in Portugal. This project used a model of Manufactures to study the process of manufacturing glass packaging. The database Project “Newglass” consists of the operating variables of the furnace and the percentage of defects found in end products of the factory model. The classification obtained through the ANFIS was compared with the results obtained in the manufacture of glass for packing. The classifications obtained in the manufacture and in the ANFIS software were also compared with the classification obtained with CART (Classification and Regression Tree).