What are the advantages of using "random forest" with "parsnip"?
I am learning parsnip from tidymodels.
The advantage of parsnip is that you only need to specify the engine and mode in a certain format, and then you can make predictions.
It also allows you to unify hyperparameter names and so on.
random forest was used as an example.
You can specify the same arguments in three different packages, ranger, sparklyr and randomForest.
I think this is a great advantage, but if randomforest has the same algorithm in these engine, can't we limit ourselves to just one of them?
In what cases do I need to use different randomforests for different engines?
One possible reason is that it is easy to work as a wrapper when the dependency randomforest is deprecated or another package comes out.
However, what are the advantages of using the current three different randomforests?