Glass classification and identification based on multiple clustering model

Abstract

Ancient glass is extremely susceptible to weathering by the influence of the buried environment, resulting in a change in the proportion of glass components. In order to classify different kinds of glass, this paper uses k-means++ clustering method to analyze the classification law, and summarizes the major and minor chemical composition data of the cluster center, so as to establish a glass classification model. In order to identify unknown types of glass, this paper mixes the data of eight groups of unknown types of glass into the known data, outputs the types of eight groups of glass through cluster analysis, calculates the center distance of the lead barium glass group to the subclass of the type in these eight groups, and then divides the subclass, and at the same time divides the subclass of the high potassium glass in the eight groups.