Analysis of Chemical Composition Content of Glass Artifacts Based on Clustering Algorithm

Abstract

The development of civilizations over the centuries has left mankind with many valuable treasures, especially the Silk Road which carried the exchange of ideas and cultures between China and the West in ancient times, and ancient glassware was born in the process. This paper first analyses the correlation between whether the artefacts are weathered and the relationship between glass type, decoration and color to find the degree of correlation; then the variance of each component before weathering is used to find the substance with the most stable change as the eigenvalue, and the K-Means clustering algorithm is used to cluster the two categories of glass artefacts, and by exploring the partial least squares regression equation for each category, multiple By exploring the partial least squares regression equation for each class, the trend of change in chemical content before and after weathering was found; finally, the pre-weathering content was predicted by determining the class of the post-weathering eigenvalue data and substituting back into the regression equation for each class.