Supervised machine learning on lending and credit risk.

This project aims to predict credit risk using various ensemble machine learning techniques. I have also tried to handle imbalance by using various sampling methods.

Credit risk is an inherently unbalanced classification problem, as the number of good loans easily outnumber the number of risky loans. I employed Machine Learning techniques to train and evaluate models with unbalanced classes. I used imbalanced-learn and scikit-learn libraries to build and evaluate models using resampling. I also evaluated the…

Using the imbalanced-learn and Scikit-learn libraries to build and evaluate machine learning models.

Use Python to build and evaluate several machine learning models to predict credit risk for FinTech firms.

A predictive model to anticipate customer churn in telecom. Using supervised ML techniques, it identifies at-risk customers based on usage patterns and service plans. Proactively retaining customers, reducing attrition costs.

Resampling exercise to predict accuracy, precision, and sensitivity in credit-loan risk

Module 12 - Using the imblearn , I'll use a logistic regression model to compare 2 versions of a dataset. First, I’ll use the original data. Next, I’ll resample the data by using RandomOverSampler. In both cases, I’ll get the count of the target classes, train a logistic regression classifier, calculate the balanced accuracy score, generate a con

Performed feature engineering, cross-validation (5 fold) on baseline and cost-sensitive (accounting for class imbalance) Decision trees and Logistic Regression models and compared performance. Used appropriate performance metrics i.e., AUC ROC, Average Precision and Balanced Accuracy. Outperformed baseline model.