Classification Of Ancient Ceramics Based On Multivariate Statistics And Intelligent Algorithms

China's ceramic culture has a long history.With the development of the modern economy,the demand for ceramic products,especially for ancient ceramic products,is rising in domestic and foreign markets.People are increasingly inclined to purchase ancient ceramics to enrich the culture and improve aesthetics.However,the antique style of porcelain has been popular since the early Song Dynasty.The purpose of imitating ancient style is generally because of the need to inherit and develop ceramic skills,or based on the personal preferences of people from all walks of life at the time,but more based on profit.Business purpose.Researchers want to identify whether ancient ceramics are fake products.Generally,it is necessary to rely on the experience accumulated in the work to compare the ancient ceramic samples to be identified with the excavated ancient ceramic cultural relics specimens.However,due to the variety of ancient ceramic products,the manufacturing processes of different regions of different dynasties are different.If the ancient ceramics are only identified based on traditional experience,there are many subjectivity and one-sidedness.Therefore,how to use the corresponding statistical algorithms for analysis based on the characteristics of the ancient ceramic sample data has become a hot issue for ancient ceramic researchers.In recent years,some research institutions have used advanced technology to study the chemical composition of ancient ceramics.The chemical composition of these ancient ceramics must contain information about the age and geography of the firing,which can be used to analyze the source and the source.This paper takes the chemical composition data of ancient ceramics from the four dynasties of the Song,Yuan,Ming and Qing Dynasties in Jingdezhen area as the research object,and uses the principal component analysis and discriminant analysis method based on K-means clustering in multivariate statistical analysis,the random forest algorithm in intelligent algorithm and The extreme learning machine algorithm and the particle swarm optimization extreme learning machine algorithm are analyzed and processed to find out the relationship and obtain the classification result.In this paper,the application of principal component analysis and discriminant analysis based on K-means clustering in the classification of ancient ceramic samples is realized by SPSS software.Different function expressions are established and classified according to this.The programming of MATLAB software is used to realize the application of random forest algorithm and extreme learning machine algorithm and particle swarm optimization limit learning machine algorithm in ancient ceramic sample classification.These statisticalmethods are compared and analyzed,and the causes,advantages and disadvantages of the differences between different methods are discussed,which provides a valuable basis for the classification of ancient ceramics.The research conclusions of this paper play a certain reference role in the identification of ancient ceramics.