header image
Home arrow Tutorial arrow Centring and scaling of data
Centring and scaling of data PDF Print E-mail
Written by Agnar Höskuldsson, DTU   
It is common to centre and scale data such that each variable in the analysis have mean zero and unit variance. Here the situation is discussed briefly. There are in the literature different views on what one should do, some views are personal experiences and others are based on theoretical considerations.
Often these views reflect misunderstanding of the situation, especially concerning the question of scaling. Here the procedures of centring and scaling are explained further. The presentation here is not complete and thus the different parts may be explained further
< Prev   Next >
Search website

Editorial flash
Svante Wold's New Year Editorial 2005
Another year has passed too quickly, and we now read January 2005 on our calendars. In 2004, good things happened and terrible, happy and sad, as usual. However, chemometrics does well, spreading deeper into biology where the data sets are larger than ever with thousands of variables in, among others, gene arrays, LC-MS profiles, and NIR and IR microscopy.
Read more...
News flash
Kick-off for Chemometrics.se
I am very pleased to announce the official kick-off for Chemometrics.se. You will recognize many "old" features such as the popular Tutorial section, but we have also added a new Editorial section, with more focus on being an Editorial, rather than a tutorial.
Read more...
Tutorial flash
Why PLS sometimes has more than one component?
I want to discuss some algebraic aspects of the popular OLS and PLS regression methods in chemometrics taken from my thesis (2001). The main point I want to make here is that single-y PLS models should only have one (predictive) component.
Read more...