Application Research In Classification Of Ceramic RAW Materials And Ancient Ceramics Based On Data Mining Technology

The development of ceramics is an important part of the history of Chinese civilization,and its historical and social values are incalculable.Nowadays,with the deepening of research work in ceramic field,classification and dating of ceramic artifacts have become one of the urgent tasks to be solved in ceramic research.In terms of traditional ceramic identification methods,the classification and dating of ceramics mostly focus on two commonly used methods: one is the traditional manual identification method,that is,to distinguish the age and ceramic category of ancient ceramics by visual touch and touch.Another method is to identify the period and category of the ceramics by extracting the chemical components of the ceramic surface based on some scientific and technological means,such as the composition of the glaze.The above two methods have their own merits and demerits.Many times they will be combined and applied in the field of ceramic research to make up for the shortcomings of the two methods.In addition,considering the wide variety,complex structure and various components of ceramic raw materials and ancient ceramics,the accuracy of traditional chemical experiment analysis is low,time consuming and energy consuming.People have begun to use many new scientific and technological means to improve the classification of ancient ceramics and the effect of dating.At present,a lot of new research achievements have been achieved in the application of data mining to the related research work in the field of ceramics.For example,artificial neural network and evolutionary computation have all achieved good experimental results in the study of ceramic classification and fracture generation.Based on the preliminary research work of many scholars,and through related depth analysis and research on chemical extraction of ancient ceramic components for data analysis,the author found out the relationship and the laws between them through the effective data analysis of the chemical composition of the extracted ceramic raw materials and ancient ceramics.Finally the methods of neural network,genetic optimization algorithm,support vector machine and random forest are applied to the classification and research process of ancient ceramics.In the third chapter of this article,the model of the ceramic raw materials is established by BP neural network algorithm.In the fourth chapter,the model of the BP neural network is optimized by genetic algorithm to classify the ceramic raw materials.The experimental results show that the classification of GA-BP Accuracy and accuracy are significantly higher than the BP neural network model.In the fifth and sixth chapters of this paper,the support vector machine and random forest to classify the ancient ceramics.The experimental results show that the prediction accuracy of random forest is higher than that of support vector machine.Although the experimental results of prediction have a lot of differences from the ideal ones,this paper aims to provide a more subsequent research in the theory and practice of scientific guidance,so as to enrich the data mining method research in the field of ancient ceramics.As a practical application method in the ancient ceramics field,This paper can also further make up for the shortcomings of the traditional methods of identification.