Solving the problem of composition and identification of ancient glass products based on classification model

Abstract

Glass has been used as ornaments or containers in ancient times, and in recent years, with the development of the world, the existence of glass has been paid more and more attention. Examining the composition and identification of ancient glass artifacts reflects the rise and fall of glass production in each dynasty, thus reflecting the face of each dynasty, reflecting the historical understanding and importance of glass materials, which is quite helpful for archaeological research. The problem of predicting unknown classes of glass artifacts and performing sensitivity analysis is solved by a classification model based on random forest. In this paper, we first define lead-barium glass as a positive class, and build classification models for weathered and unweathered glass based on the random forest algorithm and the traditional classification algorithm, and then train them separately. ROC curve Verify the prediction effect of random forest.