Rapid Miner vs R Decision Tree Classification

NatalySimth

Is there are any differences between the decision tree models generated in RapidMiner and R? Is there a comparison between them?

varunm1

Hello @NatalySimth

The decision tree mainly depends on splitting criteria. The algorithms like java implementation of CART /C4.5 kicks in Rapidminer based on your selection of criterion like Gain ratio or Gini Index. These are similar to the R implementation of CART. The criterion plays major role in deciding the DT algorithm rather than a package unless they are a modified version of the default algorithms.

I am not sure there are threads on this community that compared both settings. If you are looking to compare, be sure to set all the hyperparameters same in both environments.  You can also use the R-scripting operator in rapidminer to do the same using R.

varunm1

Hello @NatalySimth

The decision tree mainly depends on splitting criteria. The algorithms like java implementation of CART /C4.5 kicks in Rapidminer based on your selection of criterion like Gain ratio or Gini Index. These are similar to the R implementation of CART. The criterion plays major role in deciding the DT algorithm rather than a package unless they are a modified version of the default algorithms.

I am not sure there are threads on this community that compared both settings. If you are looking to compare, be sure to set all the hyperparameters same in both environments.  You can also use the R-scripting operator in rapidminer to do the same using R.