/kg_home_credit

Kaggle - Home Credit Default Risk

Primary LanguagePython

Output:

Bureau df shape: (301436, 45)
Process bureau and bureau_balance - done in 1s
Previous applications df shape: (338381, 116)
Process previous_applications - done in 3s
Pos-cash balance df shape: (337252, 12)
Process POS-CASH balance - done in 1s
Installments payments df shape: (334527, 19)
Process installments payments - done in 1s
Credit card balance df shape: (103558, 21)
Process credit card balance - done in 1s
(356251, 260)
Mix columns - done in 0s
(356251, 260)
(356251, 260)
Starting LGB. Train shape: (307507, 260), test shape: (48744, 260)
Fold 1 AUC : 0.786109
Fold 2 AUC : 0.783495
Fold 3 AUC : 0.784444
Fold 4 AUC : 0.785503
Fold 5 AUC : 0.788927
Full AUC score 0.785689
val score: [0.786108888665005, 0.783494624780713, 0.7844437626017514, 0.7855030770737955, 0.78892733350874]
train score: [0.8571782844132909, 0.8610378021998177, 0.871286024787485, 0.8544736454625311, 0.8618261458844354]
Run train with kfold - done in 4718s
Full model run - done in 4726s