YANG Yihan, LIU Bingxiang
(School of Information Engineering, Jingdezhen Ceramic Institute, Jingdezhen 333403, Jiangxi, China)

Abstract: In order to solve the problem of overwhelming classifications, complicated structures, variegated components as well as the shortages of low precision, time and labor consuming rising in the process of traditionally-used chemical experiment for ceramic raw material classification this paper has adopted the BP neural network optimized through genetic algorithm to establish a model to classify ceramic raw materials. The model has been put into the analysis and practice on 20 groups of samples and the classification results have been compared with those of 10 groups of testing samples. The results of the experiment demonstrate that the predicted value by the genetic algorithmoptimized BP neural network model is in compliance with the actual value. Such model is proved to have a high precision and practical value.

Key words: ceramic raw materials classification; BP neural network; genetic algorithm