I have several datasets, each with with hundreds of samples. I have different metadata for each data set, which contains about 50 variables per sample. Some of this metadata is clearly redundant. For example, each sample has multiple unique Id's related to different databases which I can easily remove/merge.

However, I know that other variables are also redundant. For example plot naming/numbering (it's agricultural data) overlaps with some tested variables. Is there an easy way to reduce the dimensions of my metadata? Much, but not all of it is categorical.

I was thinking that it might make sense to use Canonical Correspondence Analysis (CCA) here (similar to PCA).

I'm using julia for my project, but maybe I should do this part in r. Or does somebody know a package that can do CCA in julia?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.