A Study of Weathering and Classification Patterns of Glass Artifacts

Abstract

Background: Ancient glass is very susceptible to the influence of the buried environment and weathering, in the weathering process, the internal elements of the glass and environmental elements for a large number of exchanges, resulting in a change in the proportion of its composition and thus affecting the correct judgment of its category.This paper discusses the identification of ancient glass artifacts (high-potassium glass and lead-barium glass) in relation to the effects of weathering, the statistics and patterns of content, and the study of issues related to the classification of types.
 Methods: The data were obtained from the data set of Question C of the 2022 Gaoxueshe Cup National University Student Mathematical Modeling Competition. After eliminating the invalid data, analytical methods such as association rule analysis, factor analysis, cluster analysis, and algorithms such as Apriori and K-means++ were used for problem solving.
 Results: A total of 56 glass artifacts and 56 artifact sampling sites were included in the study, and the main findings were as follows: the relationship between the weathering of the surface of glass objects and their glass type, decoration, and color under the two-, three-, and four-correlation rules; the classification of glass objects into high-potassium or lead-barium glass based on the five compositional categories (SiO2, PbO, CaO, Fe2O3, and K2O); and the clustered dendrograms. The most homogeneous classification of high-potassium or lead-barium glass into each of the three subcategories was based on magnesium oxide and aluminum oxide, whose content varied considerably before and after the weathering process; and finally, changes in the chemical compositional correlation between the different categories were also analyzed.
 Conclusion: Solved the problems related to the identification of the composition of this batch of ancient glass artifacts in relation to the effects of weathering, the statistics and patterns of the content, and the classification of the types.