科技创新与应用2024，Vol.14Issue(7)：41-46,6.DOI:10.19981/j.CN23-1581/G3.2024.07.010
基于成分数据分析与模糊模式识别的古代玻璃种类鉴别
摘要
Abstract
Taking the composition data of a batch of ancient glass products as an example,this paper studies the application of composition data processing method and fuzzy pattern recognition in the classification and identification of ancient glass.Three logarithmic ratio transformation methods of asymmetry,symmetry and equidistance are used to transform the ancient glass composition data.In view of the fact that the zero value in the data can not be transformed by logarithmic ratio,a multiplicative substitution method which can keep the space operation of simplex is introduced.The effects of three logarithmic ratio transformation methods on fuzzy pattern recognition results are compared respectively,and random forest and support vector machine algorithms are introduced to further analyze the fuzzy pattern recognition method.Through the approximate zero value replacement,the statistical characteristics of the ancient glass composition data are basically unchanged,and the multiplication substitution method has a good replacement effect.The accuracy of fuzzy pattern recognition results and the characteristics of transformation methods show that the fuzzy pattern recognition method based on symmetric logarithmic ratio transformation is the best.The theory of composition data analysis is introduced into the classification and identification of ancient glass,which provides a scientific method for the processing of ancient glass composition data.The introduction of fuzzy pattern recognition method provides a new simple,fast and accurate classification method for archaeologists and geological researchers.关键词
古代玻璃/成分数据/零值替换/对数比/模糊模式识别Key words
ancient glass/composition data/zero substitution/logarithmic ratio/fuzzy pattern recognition分类
社会科学引用本文复制引用
王保乾,蒋剑军..基于成分数据分析与模糊模式识别的古代玻璃种类鉴别[J].科技创新与应用,2024,14(7):41-46,6.基金项目
安徽省重点研究与开发计划项目(202004a05020010) （202004a05020010）
安徽省 2020 年高校自然科学重点研究项目(KJ2020A0973) （KJ2020A0973）
安徽省级教学团队(2019jxtd105) （2019jxtd105）
安徽省精品线下开放课程(2020kfkc553) （2020kfkc553）
