polaris learn: allow user to specify user-defined hyperparameters for model
Currently, the hyperparameters for the XGBoost model we use in polaris learn
are hardcoded. We should allow users to vary these.
As discussed in #64, this may be best done through a configuration file, rather than more command line arguments. However, questions about how to do this should not block implementation of the feature; we can always switch mechanisms (CLI to config file, say) later on.
Edited by Red Boumghar