creditrisk

scikit-learn compatible tools for building credit risk acceptance models

This project mainly implements the Monotonic Optimal Binning(MOB) algorithm in SAS 9.4. We extend the application of this algorithm which can be applied to numerical and categorical data. In order to avoid the problem of creating too many bins, we optimize the p-value iteratively and provide bins size first binning, monotonicity first binning, a…

A Collection of Python Functions for Vasicek Distribution

End-to-end ML platform predicting loan default risk — data pipeline, 4-model comparison (LogReg/RF/XGBoost/LightGBM), SHAP explainability, MLflow tracking, drift monitoring, FastAPI + Streamlit, Dockerized.

Projeto Hackathon 2025: Implementação de Data Lakehouse no Databricks Free para criação de book de variáveis e modelos de previsão de inadimplência.

A Collection of R Functions for Vasicek Distribution

An end-to-end Credit Risk Project. A lender earns interest on good loans and loses money on defaults. Given an applicant's profile, estimate the probability of default (PD), turn it into a score, and choose an approval cutoff that balances approval rate against losses.

Credit card credit dataset analyzed using multiple machine learning models to determine which model best fits the data, reduces bias and predicts credit risk. Undersampling and oversampling done using various python libraries (imbalanced-learn and scikit-learn).

Full toolkit for credit risk monitoring/validation

In this project, we wanna create Credit Risk Management by using Machine Learning, so we dig into the data. what we do for the next steps are Data Preparation, EDA(Exploratory Data Analysis), Data Visualization, Data Preprocessing (Handling Outliers, Missing Value, Feature Encoding, Standardization, and Normalization), Creating Machine Learning …

End-to-end Machine Learning project for customer credit risk prediction using PostgreSQL, XGBoost, FastAPI, and GitHub.

Risk modeling: scoring and RWA estimation

Retail Credit Risk Assessment Engine

My practice notebooks from tasks on Kaggle

Credit card customer segmentation, churn prediction, and revenue analytics with Power BI dashboard

Lendora is an intelligent credit scoring platform that combines machine learning, behavioral analysis, OCR, and generative AI to assess an individual's creditworthiness. It is designed for lenders, fintechs, and microfinance institutions to streamline credit evaluation and minimize risk in real time.

Credit risk analysis with different econometric and statistic tools.

Credit risk EDA + logistic regression (class imbalance handling)

ML-powered corporate financial distress prediction using Random Forest. ROC-AUC 0.93. Flask API + interactive dashboard.

Submitted Solution to Kaggle's Home Credit Competition