Composition analysis and identification of ancient glass objects based on Random Forest

Abstract

Glass is a valuable physical evidence of our early trade exchanges, and ancient glass is susceptible to weathering by the burial environment, resulting in changes in its composition ratio. In this study, the surface weathering of glass artifacts was analyzed in relation to their glass type, decoration and color; the statistical patterns of the chemical composition content of artifact samples with and without weathering were analyzed in relation to the glass type and a Random forest (RF) model was developed to predict the chemical composition content of the artifacts before weathering based on the weathering point detection data.