Mathematical modeling for compositional analysis and identification of ancient glass objects

This paper establishes a classification model of ancient glass artifacts based on the relevant data of ancient Chinese glass artifacts using logistic regression analysis. In order to study the compositional changes and identification of glass artifacts, firstly, the relationship between the surface characteristics of glass artifacts and weathering conditions is analyzed by using the chi-square test, and the data are visualized to illustrate the statistical laws; random sampling is used to eliminate the influence of unbalanced samples, and a univariate linear regression model is fitted to predict the content of the chemical compositions before weathering. Secondly, the changes in the average chemical content of the two glass surfaces under different weathering conditions were visualized and analyzed, and the pairwise test of the chemical composition data was carried out by the rank-sum test to classify the different samples based on the K-mean clustering, and the sensitivity of the classification was analyzed. Finally, grey correlation analysis was used to quantify the degree of correlation between different chemical compositions of the same type of glass, and the variation of the degree of correlation between different types of glass was visualized and analyzed.