An investigation of the significance of the association of each attribute of ancient glass products based on cardinal distribution

Abstract

The Silk Road was a key channel for trade and cultural exchange between China and the West in ancient times, and one of the gems of the early exchange was glass products. The composition and properties of glass from the East and the West often differ greatly because of the different production processes, but the problem of glass identification has been an ancient challenge. For the analysis of the composition and origin of ancient glass products, this paper is a mathematical approach to model design, avoiding the duplication of labor into this area. This paper uses chi-square test and correspondence analysis to first determine which factors are significantly correlated with glass products, and then further determine the correlation with the factors. The use of box-line plots allows for a very visual analysis of the statistical patterns of the presence or absence of chemical composition content on the surface of artifact samples. For predicting the chemical composition content before weathering, the approximate chemical composition content was first estimated based on the average rate of change of each chemical composition before and after weathering, and then a multiple linear regression model was used to correct for the chemical composition content. It was finally concluded that the weathering of glass surface was significantly correlated with glass type, and high potassium glass was not easily differentiated, while lead-barium glass was easily weathered; the silica content changed most significantly before and after weathering of high potassium glass, with an average change rate of 25.98%, and the silica and lead oxide content changed most significantly before and after weathering of lead-barium glass, with average change rates of 29.75% and 23.77%, respectively.