Analysis of glass relics based on principal component analysis and k-means clustering

Abstract

The Silk Road was an important channel for cultural exchanges between China and the West in ancient times, of which glass was a precious item representing early trade exchanges. In China, ancient glass is generally classified into high potassium and lead-barium categories. Using data collected and classified by archaeologists for glass relics, this paper wants to investigate the differences between the classification results based on their own chemical composition data and those given empirically, and the effect of weathering on the chemical composition of glass relics. To this end, the chemical composition was first reduced using principal component analysis for non-weathered and weathered relics respectively, and three principal components were extracted for each, with variance explained at 94.49% and 98.44% respectively. Then, the number of clusters of K-means clustering was determined to be 2 after the analysis of silhouette coefficient, and then the clustering results of non-weathered and weathered glass relics were obtained by K-means clustering. For non-weathered relics, the clusters were “class Pb” and “class Pb-K”; for weathered relics, the clusters were “class Ba” and “class Pb”. Through the above study, this paper finds that glass relics can be more carefully classified according to their own chemical composition characteristics, and weathering has a great impact on the chemical composition of glass relics.