Analysis of the identification results of the ancient glass products based on the progressive regression model

Abstract

Glass is a precious witness of trade on the Silk Road. It was first introduced into China from West Asia and Egypt. Its production method was also acquired by ancient Chinese craftsmen, and local glass was created. Ancient glass is easily weathered by the influence of buried environment. In the process of weathering, the proportion of chemical composition of cultural relics will change, thus affecting the correct judgment of its category. This paper conducts the composition analysis and identification of ancient glass products by establishing the mathematical model of the correlation before and after different types of glass weathering. This paper first preprocesses the data to generate dummy variables. Then the logic and regression model are established, take the Sigmoid function as the connection function, then calculate the regression coefficient β with the maximum likelihood estimation, and then use the binary Logistic regression to compare the predicted value with the true value, to obtain the prediction success rate and the identification results. A stepwise regression model was then used to perform the sensitivity analysis of the classified results. Finally, the model is optimized, the data is imported for the Fisher linear discrimination, and the results obtained by the Fisher discrimination method are compared with the logistic regression results to make the results more accurate.