【作者】 冯文锴；

【导师】 孙柏年；

【作者基本信息】 兰州大学 ， 古生物学与地层学， 2009， 硕士

【摘要】 本文将数理的方法应用于古生物学的研究中,用聚类分析方法对新生代植物化石进行了分析。本文将化石属名和化石采集地的对应关系用二进制表示,尝试应用聚类分析中Ward’s method方法对古近纪和新近纪植物化石分别进行了聚类分析。研究结果与前人的研究较为吻合,因此,此聚类分析法可以对植物区系划分进行适当的判断,在此基础上提供新的研究方法和划分依据。本文还将聚类分析应用到对矿物原生晕的分析中,将矿物的地球化学分析数据进行统计。把矿物原生晕数据进行皮尔森相似性系数计算,运用聚类分析法中的最长距离法,得出元素聚类图,取相似系数时元素可分为不同的元素组合,并显示分类间的联系。揭示出不同元素的一些特征。初步显示不同矿物元素间的联系及联系的程度,以便为以后对矿物的成因和富集规律做初步的研究。更多还原

【Abstract】 In the present paper, clustering analysis as one of the biostatistics methods was applied to study the Cenozoic macrofossil floras of China. The genus and locations of fossils were represented by the binary system, then the Ward’s method were was used to analyze the paleophytogeograhy of China in the Paleogene and Neogene age, respectively. Our results were consistent with former studies on the geographic demarcation of the Cenozoic floras of China. So it can be concluded that the clustering analysis can be used for the demarcation of plantage, which will provide a new method and criterion for the research of paleophytogeograhy. In this paper, cluster analysis was applied to the analysis of mineral elements in the geochemical analysis of data clustering, in aim to calculate the Pearson coefficient of primary halo data. In addition, clustering map of elements can be obtained using the most long-distance method.The When we choosing an intraclass correlation coefficient can be divided into different groups to calculate the correlation coefficients in order to display the links between categories and the characteristics of different elements. The results can be helpful for us to study the relation among the different mineral elements, furthermore, to research the causes and the roles of mineral enrichment.更多还原