Timeline for my xgboost model accuracy decreases after grid search with
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 8, 2021 at 17:25 | vote | accept | sameh sharawy | ||
| May 24, 2020 at 9:04 | comment | added | German C M | you can find examples fo bayesian optimization in my former answer; about your questions, I find them also interesting to evaluate (I will try to reproduce it), but about the second one, I think it could also be because there are other hyperparameters you did not consider which, combined with another ones, give a different setting | |
| May 24, 2020 at 9:00 | history | edited | German C M | CC BY-SA 4.0 |
example of bayesian optimization with hyperopt
|
| May 21, 2020 at 22:58 | comment | added | sameh sharawy | the search.best_estimator_ gives me the default XGBoost hyperparameters combination, i have two questions here, the first, the default classifier didn't enforce regularization so could it be that the default classifier is overfitting, the second is that the grid provided already contain the hyperparameters values obtained in search.best_estimator_, why the search.best_params_ wasn't the same as search.best_estimator ?, thanks for the reply, and i am interested in an example about bayesian tuning. | |
| May 21, 2020 at 9:41 | history | edited | German C M | CC BY-SA 4.0 |
added 22 characters in body
|
| May 21, 2020 at 9:19 | history | answered | German C M | CC BY-SA 4.0 |