Is it the case that the mixOmics package includes automatic scaling and centering of features by default in its functions, including splsda
? When do I need vs not need to center and scale my data ahead of using MixOmics packages as part of the data pre-processing?
Yes, the help file (e.g ?splsda
) does state that scaling is by default in the PLS related methods. Scale means scale & center so you do not need to do anything beforehand, but make a decision when you run the method as whether you want to scale. In general we advise to leave these parameters as default (i.e scale) as you will be dealing with variables with different variances and means, and this will have an effect on the results. Just look carefully at the arguments in the help files.
![Screenshot 2024-06-14 at 8.12.34 AM](https://canada1.discourse-cdn.com/free1/uploads/mixomicsusers/original/1X/e5e75d9989ca04bf0b5525a34cbf103597e3ebe4.png)
Kim-Anh
Thank you very much for clarifying and the reference in the help files!