Research on component content model of ancient glass products based on statistical analysis

Abstract

This paper mainly analyzes and identifies ancient glass products based on the surface characteristics and chemical composition content of ancient glass products, preprocesses the given data, establishes relevant statistical models, and uses SPSS, MATLAB, PyCharm and other software to perform statistical analysis on the surface of ancient glass. Weathering chemical composition law. To classify ancient glass types, we used the BP neural network classification prediction model to train, and the classification rules obtained depended on the difference of chemical composition content, and we tested the places where there was an accuracy dispute between the two. The classification of subclasses is based on the index evaluation using the method of hierarchy-entropy weight-coefficient of variation. The chemical components of potassium oxide and lead oxide are used to subclassify different ancient glass types. According to the evaluation results, it is proved that they are reasonable, and then the classification is made. Sensitivity analysis was performed on the results, and OAT was used to adjust its parameters to achieve the best subclassification effect.