Robust Multivariate Calibration in Environmental Studies

Abstract

The main goal of this article is to present a strategy how to deal with multidimensional environmental data sets containing outliers. For calibration purposes we recommend the robust Partial Least Squares (rPLS) method, which can well describe data majority not influenced by outlying observations. This method is used to describe quantitative relationships between concentration profiles of metals measured for sediments and the parameters analyzed in water in the Saale river.

Keywords:

• Multiple outliers
• Evolution program
• Robust PLS