一个多元校正的稳健诊断新方法

Robust Diagnostics in Multivariate Calibration

提出一种新的稳健诊断方法。与最小二乘估计结合进行混合物光谱中非线性点的诊断。文中探讨了该方法的性能，用计算机数字模拟及实际多组分光谱体系对其进行检验，展示了此诊断方法在分析化学计量学中实际应用的可行性。

A new Robust diagnostic method is introduced and combined with least square regression in multivariate calibration.The advantages of this diagnostic method have been verified by using numerical simulation and the treatment of real spectrophotometric data.The feasibility of the application of robust diagnostic in chemometrics has been confirmed.