-
Notifications
You must be signed in to change notification settings - Fork 108
Open
Description
Hi, thank you for sharing your code.
I found that in tune_smote.py
the metric metrics = train_catboost(...)
.
I am wondering that we want to tune SMOTE to get better sampling data, however, the quality of sampled data is measured by the performance of catboost that means the catboost should have a tuned hyperparameters. As I understand, we can only tune SMOTE with a tuned CatBoost or we can tune CatBoost if we have a high quality sampling data from SMOTE. So if it is possible, for each SMOTE tuning iteration, we will get a data sample, then we will tune CatBoost.
for iteration (tune SMOTE):
data = sample_smote
for iteration (tune CatBoost):
tune CatBoost
I cannot find where you deploy the tune_smote.py
.
Best,
Metadata
Metadata
Assignees
Labels
No labels