Advertisement
Published in:
2014 | OriginalPaper | Chapter
6. Sliced Inverse Regression
Author : Sylvain Girard
Published in: Physical and Statistical Models for Steam Generator Clogging Diagnosis
Publisher: Springer International Publishing
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Abstract
Principal component analysis can be used as a dimension reduction method by discarding the components whose variance is below a given threshold. Projecting the model output on the low dimensional subspace thus determined preserves its most salient features. However, this only uses the information carried by the output data. As the physical model links the input to the output, one can be used to inform about the other. The sliced inverse regression (SIR) method was developed by Li (1991) to reduce the dimension of a model inputs using both an sample and the associated outputs. It was chosen here because of its robustness and ease of use: contrary to some other supervised learning methods, it does not require fitting nor parameter tuning and is applicable to a vast range of models.