基于地球化学参数的海底沉积物端元定量判识方法研究 Ⅰ.分离沉积物端元的最优斜交成分因子算法

Parting method of end-members from submarine sedimentby geochemistry data I. Algorithm of optimum oblique composition factor analysisfor parting end-members from submarine sediments

系统介绍“分离沉积物端元的最优斜交成分因子算法”，此算法包括主成分分析、因子得分成分化和最优斜交变换三部分．此算法可获得成分化的非负的斜交因子得分，弥补了Ｑ型因子分析仅能用于样品间的相互关系及样品的分类研究，而不能胜任海底沉积物端元物质分离工作之不足．方法适合于沉积物物源端元定量分离工作．将此方法应用于冲绳海槽海底表层样，成功地分离出长江陆源物质．此方法同样适合于诸如岩石、水体、土壤污染等混合过程的端元定量分离工作．

Algorithm of optimum oblique composition factor analysis is introduced. It includes three procedures: Q-factor analysis, composition transforming and optimum oblique transforming. It can derive out factors which are composition and nonnegative scores, It is not onlyadaptable to sample classification, but also to parting end-members from submarine sediments. Finally, the method is demonstrated on Oki nawa Trough submarine sediment data sets, Changjiang origin end-member is parted out from those data sets successfully. The method is alsoadaptable to petrolog, Pedology and water pollution study.