Challenge: Solving Task Using XGBoost

Tarefa

Swipe to start coding

The "Credit Scoring" dataset is commonly used for credit risk analysis and binary classification tasks. It contains information about customers and their credit applications, with the goal of predicting whether a customer's credit application will result in a good or bad credit outcome.

Your task is to solve classification task on "Credit Scoring" dataset:

Create Dmatrix objects using training and test data. Specify enable_categorical argument to use categorical features.
Train the XGBoost model using the training DMatrix object.
Set the split threshold to 0.5 for correct class detection.

Note

'objective': 'binary:logistic' parameter means that we will use logistic loss (also known as binary cross-entropy loss) as an objective function when training the XGBoost model.

Solução

import numpy as np

import xgboost as xgb

from sklearn.datasets import fetch_openml

from sklearn.model_selection import train_test_split

from sklearn.metrics import f1_score

# Load the Credit Scoring dataset

data = fetch_openml(name="credit-g", version=1, parser='auto')

X = data.data

y = data.target

# Convert target to binary (1: Good, 0: Bad)

y = (y == 'good').astype(int)

# Split the data into training and testing sets

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Create DMatrix objects for XGBoost with categorical features enabled

dtrain = xgb.DMatrix(X_train, label=y_train, enable_categorical=True)

dtest = xgb.DMatrix(X_test, label=y_test, enable_categorical=True)

# Set hyperparameters

params = {

'objective': 'binary:logistic',

}

# Train the XGBoost classifier

model = xgb.train(params, dtrain)

# Make predictions

y_pred = model.predict(dtest)

y_pred_binary = (y_pred > 0.5).astype(int)

# Calculate F1-score

f1 = f1_score(y_test, y_pred_binary)

print(f'F1-score: {f1:.4f}')

Tudo estava claro?

Obrigado pelo seu feedback!

Seção 3. Capítulo 6