基于机器学习算法和优化模型的古代玻璃制品成分分析与鉴别

Study on Component Analysis and Identification of Ancient Glass Artifacts Based on Machine Learning Algorithms and Optimization Models

文章利用机器学习算法和优化模型对古代玻璃制品的化学成分进行了分析和鉴别,并采用卡方检验分析了玻璃文物的风化情况,以及风化情况与类型、纹饰和颜色的关联性。研究结果表明,风化对化学成分含量有显著影响,尤其是S_(i)O_(2)和Na_(2)O等物质的含量变化最为明显。基于以上发现,文章建立了支持向量机的二分类模型,以区分高钾玻璃和铅钡玻璃,并进一步通过主成分分析和K-means聚类法进行亚类划分。该方法成功划分了不同的玻璃亚类,通过模型敏感性分析,证实了模型的有效性和准确性。

This article employs machine learning algorithms and optimization models to examine and identify the components of ancient glass artifacts,and adopts chi-square test to analyze the weathering of these glass relics,as well as the correlation between weathering and type,decoration and color.The research findings indicate that weathering exerts a significant impact on chemical components,especially S_(i)O_(2) and Na_(2)O.Based on the above findings,a binary classification model using support vector machines was proposed to distinguish between high potassium glass and lead-barium glass,and further divides them into subcategories by means of principal component analysis and K-means clustering.Finally,sensitivity analysis confirmed the effectiveness and accuracy of the model.