Study on Composition Analysis and Species Identification of Glass Relics Based on the Multiple Linear Regression Model

Antique glass products are highly susceptible to environmental influences and weathering, their chemical composition ratios prone change. Given this, this article is based on integrating known data processing mathematical methods such as comprehensive evaluation mean analysis establish a multiple linear regression model explore the changes in surface composition. According clustering method, accurately classify subcategories rationality sensitivity of classification results. Finally, use Euclidean distance determine unknown category cultural relics be tested. The results show that: (1) For lead barium glass, Na2O Al2O3 have protective effect weathered relics, SiO2 content decreases after while PbO, BaO, P2O5, CaO increases; high potassium SrO, SnO, SO2 almost zero, remains unchanged before weathering. increases other elements decreases. (2) successfully subdivided subclass into four categories: low SiO2, BaO-PbO-CuO, PBO BaO-SO2, BaO PbO-SiO2, SiO2-PbO BaO-Al2O3. Three types glass: SiO2-CaO-Al2O3.