An Investigation into the Composition Analysis and Identification of Ancient Glass Artifacts

Abstract

This study employs association analysis and chi-square tests to examine the relationship between weathering and glass type, color, and patterns in ancient glass artifacts. The findings reveal a stronger association between lead-barium glass and weathering, with weaker associations for color and patterns. Combining both methods demonstrates that glass type exhibits the strongest correlation with weathering, followed by color, while patterns remain unrelated. Utilizing kernel density plot analysis, the research investigates variations in chemical composition pre- and post-weathering, predicting component content before weathering. The classification of glass is based on intrinsic and extrinsic characteristics, highlighting notable differences in potassium oxide, lead oxide, and barium oxide between high-potassium and lead-barium glass. Scatter plots further illustrate these compositional disparities. Applying a decision tree, glass is classified according to patterns, color, and surface weathering, determining the influence of various factors. Hierarchical clustering is then used to categorize chemical compositions, identifying five categories for lead-barium glass and nine for high-potassium glass.