Comparison and Analysis of Accuracy of Various Machine Learning Algorithms in Ancient Glass Classification

Glass is one of the earliest man-made materials invented by mankind, and its appearance and use have been used in human life for thousands of years. A large part of the glass unearthed in ancient China is glass with high potassium content and lead barium glass, that is, high potassium glass and lead barium glass, which have completely different storage environments. In order to make different kinds of ancient glass intact in storage without corrosion, the classification of high potassium and lead barium glass is particularly important. This paper aims to analyze the classification principles of each machine learning classifier and compare the classification performance differences of each classifier. We selected the latest competition dataset, compared the advantages and disadvantages of various machine learning methods in glass classification and came to a conclusion, for the classification of ancient glass, this paper mainly did the following aspects of work: one is the selection of the dataset and the preprocessing of text information, the other is to divide the data into two parts, one part for training, one part for testing, with each machine to train the model, with the test set to test the model. Finally, parameters such as accuracy, precision, recall, and F1 score are calculated according to the email classification results, and the models are compared. The results show that the adaboost classification model and XGBoost model have the best classification effect, and they can obtain better generalization ability in the test data, which should be more applied in practice.