Pipeline riesgo de crédito#
import pandas as pd
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import (
BaggingClassifier,
RandomForestClassifier,
AdaBoostClassifier,
GradientBoostingClassifier,
StackingClassifier,
)
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report
import xgboost as xgb
# Cargar el archivo CSV (ya lo hemos cargado y limpiado previamente)
credit_risk_data = pd.read_csv("credit_risk_data.csv")
credit_risk_data = credit_risk_data.drop(columns=["ID"])
# Dividir los datos en características (X) y etiqueta (y)
X = credit_risk_data.drop(columns=["Estado del Préstamo"])
y = credit_risk_data["Estado del Préstamo"]
# Dividir el conjunto de datos en entrenamiento y prueba
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=34)
Creación de los modelos:#
Pipeline()
models = {
"Logistic Regression": Pipeline(
[
("scaler", StandardScaler()),
("classifier", LogisticRegression(max_iter=1000)),
]
),
"SVM": Pipeline(
[
("scaler", StandardScaler()),
("classifier", SVC(kernel="rbf", probability=True)),
]
),
"Decision Tree": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
DecisionTreeClassifier(
max_depth=3, min_samples_split=5, min_samples_leaf=2
),
),
]
),
"Bagging": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
BaggingClassifier(
base_estimator=DecisionTreeClassifier(),
n_estimators=100,
max_samples=0.8,
bootstrap=True,
n_jobs=-1,
random_state=42,
),
),
]
),
"Random Forest": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
RandomForestClassifier(
n_estimators=100,
max_depth=3,
max_features="sqrt",
random_state=42,
n_jobs=-1,
),
),
]
),
"AdaBoost": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
AdaBoostClassifier(
base_estimator=DecisionTreeClassifier(max_depth=1),
n_estimators=100,
learning_rate=0.1,
random_state=42,
),
),
]
),
"Gradient Boosting": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
GradientBoostingClassifier(
n_estimators=100, max_depth=3, learning_rate=0.1, random_state=42
),
),
]
),
"XGBoost": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
xgb.XGBClassifier(
n_estimators=100,
max_depth=3,
learning_rate=0.1,
subsample=0.8,
colsample_bytree=0.8,
random_state=42,
),
),
]
),
"Stacking": Pipeline(
[
("scaler", StandardScaler()),
(
"classifier",
StackingClassifier(
estimators=[
("svc", SVC(probability=True)),
(
"rf",
RandomForestClassifier(n_estimators=100, random_state=42),
),
("dt", DecisionTreeClassifier(random_state=42)),
("log_reg", LogisticRegression(max_iter=1000)),
],
final_estimator=LogisticRegression(),
cv=5,
),
),
]
),
}
# Evaluar cada modelo
accuracies = {}
for name, pipeline in models.items():
# Validación cruzada para obtener la media de la precisión
cv_scores = cross_val_score(pipeline, X_train, y_train, cv=10, scoring="accuracy")
mean_cv_score = cv_scores.mean()
# Entrenar el modelo y predecir en el conjunto de prueba
pipeline.fit(X_train, y_train)
y_pred = pipeline.predict(X_test)
# Calcular la precisión en el conjunto de prueba
test_accuracy = accuracy_score(y_test, y_pred)
# Guardar las precisiones
accuracies[name] = {
"CV Accuracy": mean_cv_score,
"Test Accuracy": test_accuracy,
"Confusion Matrix": confusion_matrix(y_test, y_pred),
"Classification Report": classification_report(y_test, y_pred),
}
# Mostrar los resultados
for model_name, metrics in accuracies.items():
print(f"Model: {model_name}")
print(f"Cross-Validation Accuracy: {metrics['CV Accuracy']:.2f}")
print(f"Test Accuracy: {metrics['Test Accuracy']:.2f}")
print("Confusion Matrix:")
print(metrics["Confusion Matrix"])
print("Classification Report:")
print(metrics["Classification Report"])
print("\n" + "-" * 40 + "\n")
Model: Logistic Regression
Cross-Validation Accuracy: 0.88
Test Accuracy: 0.87
Confusion Matrix:
[[ 46 27]
[ 12 215]]
Classification Report:
precision recall f1-score support
0 0.79 0.63 0.70 73
1 0.89 0.95 0.92 227
accuracy 0.87 300
macro avg 0.84 0.79 0.81 300
weighted avg 0.87 0.87 0.86 300
----------------------------------------
Model: SVM
Cross-Validation Accuracy: 0.94
Test Accuracy: 0.93
Confusion Matrix:
[[ 57 16]
[ 6 221]]
Classification Report:
precision recall f1-score support
0 0.90 0.78 0.84 73
1 0.93 0.97 0.95 227
accuracy 0.93 300
macro avg 0.92 0.88 0.90 300
weighted avg 0.93 0.93 0.92 300
----------------------------------------
Model: Decision Tree
Cross-Validation Accuracy: 0.92
Test Accuracy: 0.92
Confusion Matrix:
[[ 55 18]
[ 7 220]]
Classification Report:
precision recall f1-score support
0 0.89 0.75 0.81 73
1 0.92 0.97 0.95 227
accuracy 0.92 300
macro avg 0.91 0.86 0.88 300
weighted avg 0.92 0.92 0.91 300
----------------------------------------
Model: Bagging
Cross-Validation Accuracy: 0.98
Test Accuracy: 0.98
Confusion Matrix:
[[ 70 3]
[ 3 224]]
Classification Report:
precision recall f1-score support
0 0.96 0.96 0.96 73
1 0.99 0.99 0.99 227
accuracy 0.98 300
macro avg 0.97 0.97 0.97 300
weighted avg 0.98 0.98 0.98 300
----------------------------------------
Model: Random Forest
Cross-Validation Accuracy: 0.88
Test Accuracy: 0.86
Confusion Matrix:
[[ 30 43]
[ 0 227]]
Classification Report:
precision recall f1-score support
0 1.00 0.41 0.58 73
1 0.84 1.00 0.91 227
accuracy 0.86 300
macro avg 0.92 0.71 0.75 300
weighted avg 0.88 0.86 0.83 300
----------------------------------------
Model: AdaBoost
Cross-Validation Accuracy: 0.98
Test Accuracy: 0.97
Confusion Matrix:
[[ 67 6]
[ 2 225]]
Classification Report:
precision recall f1-score support
0 0.97 0.92 0.94 73
1 0.97 0.99 0.98 227
accuracy 0.97 300
macro avg 0.97 0.95 0.96 300
weighted avg 0.97 0.97 0.97 300
----------------------------------------
Model: Gradient Boosting
Cross-Validation Accuracy: 0.98
Test Accuracy: 0.99
Confusion Matrix:
[[ 71 2]
[ 2 225]]
Classification Report:
precision recall f1-score support
0 0.97 0.97 0.97 73
1 0.99 0.99 0.99 227
accuracy 0.99 300
macro avg 0.98 0.98 0.98 300
weighted avg 0.99 0.99 0.99 300
----------------------------------------
Model: XGBoost
Cross-Validation Accuracy: 0.99
Test Accuracy: 0.99
Confusion Matrix:
[[ 71 2]
[ 1 226]]
Classification Report:
precision recall f1-score support
0 0.99 0.97 0.98 73
1 0.99 1.00 0.99 227
accuracy 0.99 300
macro avg 0.99 0.98 0.99 300
weighted avg 0.99 0.99 0.99 300
----------------------------------------
Model: Stacking
Cross-Validation Accuracy: 0.98
Test Accuracy: 0.99
Confusion Matrix:
[[ 71 2]
[ 2 225]]
Classification Report:
precision recall f1-score support
0 0.97 0.97 0.97 73
1 0.99 0.99 0.99 227
accuracy 0.99 300
macro avg 0.98 0.98 0.98 300
weighted avg 0.99 0.99 0.99 300
----------------------------------------
Validación cruzada:
Se utiliza cross_val_score para realizar validación cruzada con 10
divisiones (cv=10) en el conjunto de entrenamiento. Esto proporciona
una medida de la precisión promedio (mean_cv_score) obtenida en
múltiples particiones de los datos.
La validación cruzada es útil para evaluar la estabilidad y la generalización del modelo, asegurando que no se sobreajuste a una sola división de los datos.