Analysis and identification of ancient glass components based on qualitative-quantitative data mining

Abstract

First of all, the data is visualized by grouping the data in the form and drawing a scatter diagram. By observing the distribution of the data in the scatter plot, the classification indexes of high-potassium glass and lead-barium glass were found. Secondly, the Spearman coefficient is used to test these classification indicators. The results show that the constructed classification system is more accurate, the Spearman coefficient is small, and the fitting effect is excellent. Secondly, through the combination of qualitative and quantitative methods, the cultural relics sub-category classification index was double screened, and the chemical composition of the sub-category classification index was determined by combining the law analyzed from the scatter plot and the index weight solved by the entropy weight method. Then, the high-potassium glass was divided into low-calcium and low-copper group and high-calcium and high-copper group by K-means method. The glass was divided into high copper group, low sodium and high calcium group and high sodium and low calcium group. Finally, all the unknown types of cultural relics in the table are divided into subclasses, and the sensitivity and rationality of the classification system of major classes and subclasses are tested one by one by fitting the known data and calculating the contour coefficient. It is concluded that the classification system has a good fit to the data.