古代玻璃制品的成分分析与鉴别问题的数学模型  

A Mathematical Model for Composition Analysis and Identification of Ancient Glass Products

作　　者：杨磊[1] 魏彤 宗毅恒 YANG Lei;WEI Tong;ZONG Yiheng(Public Education Department,Dalian University of Finance and Economics,Dalian 116000,China)

机构地区：[1]大连财经学院公共教学部,大连116000

出　　处：《北京印刷学院学报》2023年第9期61-66,共6页Journal of Beijing Institute of Graphic Communication

基　　金：2021年度辽宁省教育科学“十四五”规划项目(JG21DB024);2022年度辽宁省教育厅高校基本科研项目(LJKMR20221952)研究成果。

摘　　要：古代玻璃极易受埋藏环境的影响而风化,考古工作者依据古代玻璃文物样品的化学成分和其他检测手段将其分为高钾玻璃和铅钡玻璃两种类型。采用卡方检验的方法,对玻璃文物的表面风化与其类型和颜色的关系进行差异性分析,通过训练集数据来建立支持向量机回归(SVR)模型,预测风化前玻璃文物的化学成分含量。运用层次聚类算法对该两种玻璃类型的相关数据进行聚类,根据每个类别所对应的化学成分含量在风化前后的变化情况进一步进行亚类划分,然后对分类结果给出灵敏度分析。Ancient glass is very susceptible to the influence of burial environment and weathering,archeologists according to the chemical composition of ancient glass relics samples and other detection methods,it is divided into high potassium glass and lead barium glass two types.In this paper,Chisquare test method is used to analyze the difference between the surface weathering of glass relics and its type and color,and support vector machine regression(SVR)model is established through training set data to predict the chemical composition content of glass relics before weathering.The hierarchical clustering algorithm was used to cluster the relevant data of the two types of glass,and the subclasses were further divided according to the changes of the corresponding chemical components before and after weathering,and then the sensitivity analysis was given for the classification results.

关 键 词：卡方检验 差异性分析 支持向量机回归(SVR)模型 层次聚类算法 

分 类 号：TS93[轻工技术与工程]