Abstract

This paper studies the chemical composition of two major types of glass products in ancient China, and the constituent structure of residual substances in a long-term weathering environment, and proposes a classification method based on composition analysis. When researchers provide unknown categories of ancient glass samples, this results can be used to make type judgment based on their chemical composition and assist generation and other work. Based on the quantitative analysis of the chemical composition of different sampling sites, the glass sample classification model is tested by using the bp neural network and the decision tree classification model and ID3 mechanism. When analyzing the problem, we found that some of the data in the attachment did not meet the requirements of the problem setting, and even had missing values. Therefore, we cleaned the data, completed the null value and eliminated the wrong data with 0, which was embodied in lines 15 and 17 of Table 2.Using Kappa consistency matrix analysis, it is an obvious connection between the surface weathering of glass cultural relics and their types, decoration, color and other characteristics. The discrete dot plot for pooling analysis was drawn using data characteristics of surface chemical composition changes before and after weathering. To predict the chemical composition after weathering, the prediction formula was established using vector iteration. Later, we used the vector iteration method to establish the weighted average ratio prediction formula to predict the chemical composition of the glass samples after weathering

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call