Glass Classification and Identification Based on Random Forest Model and Systematic Clustering

Abstract

A large exchange of material elements occurs during glass weathering, which will affect the identification and judgment of ancient glass. To this end, the data is first used to give the basis for dividing lead barium and high potassium glass, and the difference between the chemical components of each type of glass can be compared to obtain a conclusion. Since the final number of subclasses is unknown, this paper uses systematic cluster analysis to obtain the number of subclasses and classification results at the same time. In order to identify the glass types of unknown cultural relics, this paper uses the existing data to establish a random forest model and construct a method and standard for dividing glass types.