论文部分内容阅读
采用局部加权、随机矩阵法等步骤，使建立的校准模型能充分反映未知样品的数字结构特征，达到优化的目的，从而改善偏最小二乘法在处理实际未知样品数据时的解析能力，以提高计算结果的精度。用该优化偏最小二乘法对模拟地质样品中痕量贵金属Ｒｈ、Ｉｒ、Ｐｄ进行多组分光度测定，其分析结果的相对误差均小于１０％；分别测定９．５μｇ／Ｌ、３５．０μｇ／Ｌ、６６．０μｇ／Ｌ含量的Ｒｈ、Ｉｒ和Ｐｄ７次，其标准偏差Ｒｈ为０．２；Ｉｒ为１．１；Ｐｄ为２．０。
The method of local weighting and stochastic matrix method is applied to make the calibration model fully reflect the digital structural features of unknown samples and achieve the purpose of optimization so as to improve the analytical ability of partial least square method in dealing with actual unknown sample data to improve calculation The accuracy of the result. The optimized partial least squares method was applied to the determination of trace amounts of precious metals Rh, Ir and Pd in simulated geological samples by multi-component spectrophotometry. The relative errors of the analytical results were less than 10%. The relative error of 9.5μg / L, 35.0μg / L and 66.0 μg / L of Rh, Ir and Pd seven times with a standard deviation of Rh of 0.2; Ir of 1.1 and Pd of 2.0.