0
$\begingroup$

Is xgboost with n_estimators = 100 and learning_rate = 0.1, same as xgboost with n_estimators = 50 and learning_rate = 0.2 ?

$\endgroup$
1
$\begingroup$

No, they won't have neither the same performance nor the same architecture if you were to try to visualize it. An XGBoost with 100 n_estimators and a learning rate of 0.1 is a 100 trees grown sequentially, and each tree's output is multiplied by 0.1. The latter is a model of 50 trees grown the same way and the outputs are multiplied by 0.2, so your second model ( given the usual size of datasets nowadays ) will probably underfit the data, a low number of trees to learn from the data and a high learning rate which will make it somewhat innaccurate.

Unless you're looking for another kind of similarity, this should explain it enough. I hope. If not, feel free to comment to clarify your question more.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.