Identification of glass products based on cluster analysis and random forest model

Abstract

The main chemical composition of glass is silica (SiO2). Due to the high melting point of pure quartz sand in the raw material of glass production, flux should be added in order to reduce the melting temperature. The main chemical composition varies with the flux added. In this paper, through the grey correlation analysis can be concluded that the key factors influencing the classification to the content of lead oxide, and barium oxide, and through the random forest, hierarchical clustering, and K - Means method to classify different types of glass and the division of the class, the glass can be divided into high potassium, high potassium not weathering, lead, barium, lead, barium not weathering four categories of two classes, a total of eight classes, Through the test of classification sensitivity and reliability, it shows that the three models are reasonable and can be applied.