Partial least squares with outlier detection in spectral analysis: A tool to predict gasoline properties

The aim of this study is to propose a novel partial least squares with outlier detection (PLS_OD) calibration method and show its usefulness in calibration successfully with data containing outlying objects. We apply this method in gasoline spectral analysis to predict gasoline properties. In particular, a comparative study of PLS_OD and other five methods is presented. The performances of the proposed method are illustrated on spectral data set with and without outliers. The obtained results suggest that the proposed method can be used for constructing satisfactory gasoline prediction model whether there are some outliers or not.