论文部分内容阅读
本文报道了在X-射线荧光分析多变元体系因素分析中,应用交互有效法选择体系主组份的研究。本方法在基本参数法的基础上,运用交互有效法,通过最小预报残差平方和(PRESS)的计算,对体系进行了主组份分析,确定了该体系中的最佳维数和主要影响因素。选择体系中的主组份进行回归分析,剔去了噪声,模型稳定性增强,XRFA的预报能力和分析结果准确度有了显著的提高。文中给出了分析地质样品中十三个主、次量元素的实验结果和数据对照。
In this paper, we report the selection of the main components of the system by the interactive method in the analysis of multivariate elements of X-ray fluorescence analysis. Based on the basic parameter method, the method is based on the interaction method, and the principal components of the system are analyzed by the minimum forecast residual squared error (PRESS). The optimal dimension and the main influence in the system are determined factor. The main components in the selection system are analyzed by regression analysis, the noise is removed, the stability of the model is enhanced, the forecasting ability of XRFA and the accuracy of the analysis results have been significantly improved. In this paper, the experimental results and data comparison of thirteen primary and secondary elements in geological samples are given.