Analysis of chemical composition of ancient glass products based on random forest

Due to their age, ancient glass objects are susceptible to weathering due to their burial environment, resulting in changes in their chemical composition, which can often cause significant problems for the study of ancient glass objects. Now in this paper, in order to be able to scientifically analyse and identify the composition of ancient glass artefacts, a random forest-based model for the analysis of the chemical composition of ancient glass artefacts is developed using a variety of machine learning methods and statistical learning methods. It is concluded that surface weathering of glass artefacts is not significantly correlated with the decorative properties of glass objects, but is significantly correlated with type. Furthermore, the model can well identify the type of ancient glass artefacts and predict the chemical composition of ancient glass artefacts before weathering.