Research on glass relics based on neural network and K-means clustering algorithm

Abstract

In this paper, we construct a prediction and classification model based on neural network and K-means clustering algorithm to complete the prediction of glass composition before weathering and the classification of unknown glass. In the process of studying glass relics, we discussed the important chemical components and analyzed their influence on the properties of glass. At the same time, through the collected data, the neural network algorithm is used to train and test the data to analyze the composition content of glass before and after weathering. Then, the classification simulation of the data is carried out according to the K-means clustering algorithm. Finally, the classification results of glass relics with different chemical composition in the classification model are analyzed. We conclude that there are six classification systems for high-potassium glass and lead-barium glass. Effective classification can not only reduce the difficulty of scholars' archaeological classification, but also effectively improve the research value of glass relics.