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
Hi, thank you for sharing your code.
I found that in
tune_smote.pythe metricmetrics = 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.I cannot find where you deploy the
tune_smote.py.Best,