I'm trying to optimize the parameters learning rate and max_depth of a XGB regression model:
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import cross_val_score
from xgboost import XGBRegressor
param_grid = [
# trying learning rates from 0.01 to 0.2
{'eta ':[0.01, 0.05, 0.1, 0.2]},
# and max depth from 4 to 10
{'max_depth': [4, 6, 8, 10]}
]
xgb_model = XGBRegressor(random_state = 0)
grid_search = GridSearchCV(xgb_model, param_grid, cv=5,
scoring='neg_root_mean_squared_error',
return_train_score=True)
grid_search.fit(final_OH_X_train_scaled, y_train)
final_OH_X_train_scaled
is the training dataset that contains only numerical features.
y_train
is the training label - also numerical.
This is returning the error:
FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan.
I've seen other similar questions, but couldn't find an answer yet.
Also tried with:
param_grid = [
# trying learning rates from 0.01 to 0.2
# and max depth from 4 to 10
{'eta ': [0.01, 0.05, 0.1, 0.2], 'max_depth': [4, 6, 8, 10]}
]
But it generates the same error.
EDIT: Here's a sample of the data:
final_OH_X_train_scaled.head()
y_train.head()
EDIT2:
The data sample may be retrieved with:
final_OH_X_train_scaled = pd.DataFrame([[0.540617 ,1.204666 ,1.670791 ,-0.445424 ,-0.890944 ,-0.491098 ,0.094999 ,1.522411 ,-0.247443 ,-0.559572 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,1.0 ,0.0 ,0.0],
[0.117467 ,-2.351903 ,0.718969 ,-0.119721 ,-0.874705 ,-0.530832 ,-1.385230 ,2.126612 ,-0.947731 ,-0.156967 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,1.0 ,0.0 ,0.0 ,0.0 ,0.0],
[0.901138 ,-0.208256 ,-0.019134 ,0.265250 ,-0.889128 ,-0.467753 ,0.169306 ,-0.973256 ,0.056164 ,-0.671978 , 0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,1.0 ,0.0 ,0.0],
[2.074639 ,0.100602 ,-1.645121 ,0.929598 ,0.811911 ,1.364560 ,0.337242 ,0.435187 ,-0.388075 ,1.279959 , 0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,1.0],
[2.198099 ,-0.496254 ,-0.917933 ,-1.418407 ,-0.975889 ,1.044495 ,0.254181 ,1.335285 ,2.079415 ,2.071974 , 0.0 ,0.0 ,0.0 ,0.0 ,0.0 ,1.0 ,0.0 ,0.0 ,0.0 ,0.0]],
columns=['cont0' ,'cont1' ,'cont2' ,'cont3' ,'cont4' ,'cont5' ,'cont6' ,'cont7' ,'cont8' ,'cont9' ,'31' ,'32' ,'33' ,'34' ,'35' ,'36' ,'37' ,'38' ,'39' ,'40'])
I was able to reproduce the problem and the code fails to fit because there is an extra space in your eta
parameter! Instead of this:
{'eta ':[0.01, 0.05, 0.1, 0.2]},...
Change it to this:
{'eta':[0.01, 0.05, 0.1, 0.2]},...
The error message was unfortunately not very helpful.