I have a Multivariate Time Series dataset that the variables are all categorical, but they can be expressed in order (e.g. 0=low, 1=medium, 2=high).
I want to perform feature selection on this dataset, and found a few methods using PCA, Recursive Feature Elimination and SVM, but to my knowledge they all work with continuous variables.
Is there a problem to use this feature selection algorithms in my ordered categorical variables? Or should I be using another method specifically for categorical variables in a Multivariate Time Series dataset?