metadata
library_name: scikit-learn
tags:
- random-forest
- classification
- uber
- machine-learning
pipeline_tag: text-classification
RandomForestClassifier — Uber Completion (Optimized)
Modelo supervisado para predecir el EstadoUber (0=Cancelado, 1=Completado)
derivado de Booking Status.
Métricas (test)
- Accuracy: 0.9549
- AUC-ROC: 0.9893
- Matriz de confusión: ver
metrics.json
Hiperparámetros (GridSearch)
{ "max_depth": 15, "max_features": "sqrt", "min_samples_leaf": 4, "min_samples_split": 2, "n_estimators": 15 }
Características usadas
[ "Vehicle_Bike", "Vehicle_Go Mini", "Vehicle_Go Sedan", "Vehicle_Premier Sedan", "Vehicle_Uber XL", "Vehicle_eBike", "Pay_Credit Card", "Pay_Debit Card", "Pay_NoPayment", "Pay_UPI", "Pay_Uber Wallet", "Ride Distance", "Booking Value", "Avg CTAT" ]