Analysis and Prediction of Chemical Composition of Ancient Glass Relics

Abstract

Glass was introduced into China long ago via the Silk Road. After the ancient glass was buried, its internal elements would change under the influence of the environment. It is of great significance for archaeological work to study the changes of these chemical components. In this paper, the chemical composition of ancient glass products is analyzed and predicted by using R-type clustering, BP neural network, multiple linear regression and other models.First, after using SPSS statistical data, Pearson chi square test was conducted to draw a conclusion that the category, decoration and color were not related to whether the surface of the cultural relics was weathered. Then, descriptive statistics are made on the data. Finally, the multiple linear regression model is used to predict the unknown type of cultural relics. The results show that after the weathering of cultural relics, the content of potassium oxide, calcium oxide, aluminum oxide and magnesium oxide decreases compared with that before weathering, while the content of iron oxide increases.The main chemical components of high potassium glass relics and lead barium glass relics are calculated by entropy method. Compared with silicon dioxide and aluminum oxide, they have less influence on other chemical components. Therefore, these two chemical components are eliminated to classify cultural relics. Then, the R-type clustering model is established to classify the cultural relics, and the high potassium weathering type is obtained: 22 cultural relics sampling points form a category, and the other cultural relics form a category.