Comparative Analysis of Machine Learning Models For Predicting Default in Home Credit Companies

View Full Paper

Ask AI

Bookmark

Authors

Discussion

This analysis compares machine learning algorithms in predicting default rates for credit scoring, suggesting XGBoost outperforms others.

The XGBoost model achieved an accuracy of 84.8%, outperforming logistic regression and random forest.
Performance metrics for the best model include precision at 84.9%, recall at 84.7%, and AUC at 92.3%.
Model preparation involved variable selection, missing value imputation, and transformation of categorical variables.
Feature assessments were based on principles like capacity, capital, collateral, and condition relevant to creditworthiness.

Alexander et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark