Continuous and categorical matrix for Y

Analysis
1

Hi there,

I recently performed PLS-R on a continuous matrix that was set as Y. It sounds like I will be receiving categorical data in the near future.

Is there a method one could use for multivariate analysis, similar to PLS2, where Y contains both continuous and categorical variables? Or would it be necessary to perform separate analyses for each data type?

Thank you!

2

Hi @nabo,

I am not sure if I am sure exactly what data you have, if you have:

  • two data blocks with continuous data and one categorical variables → you could consider DIABLO
  • two data blocks, one which is continuous data and the other which has both continuous and categorical variables → you will need to convert your categorical variables into a ‘dummy matrix’ (see this related post and this one) and run PLS like you have done already, but the interpretation might not be easy. Alternatively perform separate analyses as you suggested :slight_smile:

Hope that helps!
Cheers,
Eva

3

Hi Eva,

I apologize for my question being unclear.

So, I will have two data blocks. The X block is continuous, and the Y block was previously continuous with one variable. I will be getting a new categorical variable soon and this will be added to the Y block. In the end the Y block will be two variables, one continuous (the previous variable) and one categorical. It sounds like this case is related to your second bullet point?

Thanks again!

Nathan