Classification and subclassification model of glass products based on CART decision tree algorithm

Ancient glass is extremely susceptible to environmental influences and thus weathering, and after the original elements of the glass are exchanged with the environmental elements, the original chemical composition of the glass will also change and thus pose an obstacle to the determination of the type of glass (high potassium or lead-barium). In this paper, firstly, independent samples MannWhitney U test is conducted on the data. In this paper, the independent samples MannWhitney U-test is performed on the data to obtain the chemical compositions of high-potassium and lead-barium with significant differences; and then the CART decision tree algorithm is used to realize the classification of high-potassium and lead-barium glasses. Classification rule: Distinguish high potassium and lead-barium glasses by lead oxide, barium oxide and silica. For the subclassification problem. We choose suitable chemical components for each type of glass and perform K-means clustering, and find that lead oxide and silicon dioxide can be used to classify lead-barium glass into subclasses, and potassium oxide and calcium oxide can be used to classify high-potassium glass into subclasses.