Hi all,
I’m very new to this analysis so appreciate some help here!
I have a dataset with a series of continuous measures as independent variables and a series of ordinal variables (0-5) as dependent variables. May I ask if I can utilize PLS-DA to look at all the dependent variables at once (similar to what can be done with normal PLS regression/correlation), or must I look at each of the dependent variable separately in separate PLS-DA?
Much thanks!
Sophia
What is the context of your data, how many samples, what type of data? Often PLSDA is used to deal with high correlation and high dimensionality. Is your goal feature selection andor prediction?
If you have a reasonable number of samples (maybe n>30 for example), there are entire packages such as ordinalNet for Elastic Net which can do variable selection and handle correlation really well while modelling an ordinal response through an appropriate link function (it’s a form of penalized regression).
Just a thought. I am not sure how well PLSDA would work with an ordinal response. I know there are altnerative modelling approaches that maybe useful in your scenario.
Thanks @Neystale
I confirm what @Neystale said (thanks!). PLS-DA is not great with ordinal data. For an ordinal response, see this discussion here: .
For ordinal predictors: Is data suitable for mixomics? - #2 by kimanh.lecao. You could look at all of them at the same time, but they would be considered as continuous.
None of the options discussed above are great workaround though!
Kim-Anh