Practical approach to outlier detection using support vector regression

Abstract

For precise estimation with soft sensors, it is necessary to remove outliers from the measured raw data before constructing the model. Conventionally, visualization and maximum residual error have been used for outlier detection, but they often fail to detect outliers for nonlinear function with multidimensional input. In this paper we propose a practical approach to outlier detection using Support Vector Regression, which reduces computational cost and defines outlier threshold appropriately. We apply this approach to both test and industrial datasets for validation. © 2009 Springer Berlin Heidelberg.