Composition analysis and identification of ancient glass products based on population forest-KNN algorithm

Abstract

This paper is based on the KNN and Random Forest algorithms in machine learning to analyze the differences and correlations in the compositions of ancient glass objects, determine their classification and identify the unknown categories of artifacts. The classification rules of different types of glass are analyzed; for each category, appropriate sub-classification processes are performed according to their chemical composition, specific classification methods and processing results are given, and sensitivity and rationality analyses are performed on the results. The chemical composition of glass artifacts of unknown categories is analyzed, and then their categories are identified and sensitivity analysis is performed on their classification results. Samples of glass artifacts from different categories and analyze the relationships between their chemical compositions, and compare the differences that exist in the relationships between the chemical compositions of different categories.