Based on the chemical composition data of ancient glass artifacts,the chemical components of highpotassium glass and leadbarium glass were studied with Principal Component Analysis(PCA)and Grey Relational Analysis.After the chemical components highly correlated to both types of glass artifacts were found to be strontium oxide and iron oxide,with correlation coefficients of 1.0 and 1.2,respectively,the algorithmic approach of decision trees was used to analyze the chemical data associated with ancient glass,which then informed the formulation of a Random Forest model.Ultimately,artifacts numbered A1,A6,and A7 were classified as highpotassium glass,while artifacts numbered A2,A3,A4,A5,and A8 were classified as leadbarium glass.The verification of these results showed that the information loss was less than 1％,with a Brier Score of 0.0078 for the classification results.This indicates that the newly established Random Forest mode,confirmed to be very reliable,provides a basis for archaeological researchers to study ancient glass artifacts.
维普
万方数据