I am interested in predicting a continuous variable reflecting vegetative production using a collection of land use categorical variables. The dataset is a pixel-level dataset, where each pixel has a value of vegetative production and a land use categorization. There are 50+ land use categories and I want to reduce this to a smaller number. I'd like to perform dimension reduction on the land use categories based on their contribution to overall variance in vegetative production. As far as I can tell, MCA won't work unless I bin vegetative production and make it categorical. Any thoughts on how to perform dimension reduction on categorical variables based on values of a continuous variable would be greatly appreciated.
To be clear - I'm interested in the construction of the matrix required and the package to process the matrix either in R or Python. Thanks!