Dealing with Correlated Data

  • Herbert RueferEmail author


There is no interaction between parameters regarding systems designed for a technical purpose, or the interaction is well known and can be decoupled. In contrast, chains of chemical reactions which are controlled by enzymes with feedback mechanisms exist in the field of biology with an emphasis on humans. To analyze such a system by assigning variables to an orthogonal array can be misleading due to the numerous interactions which confound the parameter effect. Thus, the first step necessary is not to analyze parameters to obtaining the system response but to recognize the appearance of objects or patients or in a more generic way, a pattern. If slightly different patterns can be distinguished, the identification of parameters responsible for a specific pattern needs to be the second step which must be taken. The first step called pattern recognition, invented by P.C. Mahalanobis, considers all paired parameter correlations. Regarding the second step, orthogonal arrays and SNR indices are applicable for further analysis. By means of the SNR key figures, the individual parameter effect contribution to a specific pattern is thus investigated known as root-cause-analysis.


Orthogonal arrayOrthogonal Array interactionInteraction Correlation coefficientCorrelation Coefficient Inverse Correlation Matrix Mahalanobis Distance 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.BurghausenGermany

Personalised recommendations