We can consider two cases. Fitting 2 RandomForestClassifiers 500 trees each and average their predicted probabilities on the test set or we can fit a RandomForestClassifier with 1000 trees and use it to get test set probabilities. All hyperparameters are the same for all models.
what could be the possible variations in results for both cases !!!
I guess the result will be close for both cases as in random forest every tree is independent.
Hi
Please check the following work. It describes the process in details. I believe it will help you.
https://medium.com/better-programming/how-to-divide-data-into-train-and-test-assuring-representativeness-c4c12c215d79
Regards
Have a look at https://towardsdatascience.com/an-implementation-and-explanation-of-the-random-forest-in-python-77bf308a9b76
Depending upon data, there can be a (slight) difference in recall/accuracy.
- Badges
- Science topic