Customer Churn Prediction using Explainable Machine Learning

This repository includes an end-to-end machine learning project that uses Logistic Regression to forecast customer churn and SHAP to explain model decisions.

Overview of the Project

Businesses can identify clients who are likely to discontinue using a service by employing customer churn prediction. Customers are classified as either churned or not in this research using a supervised learning approach based on billing, service usage, and demographic data.

DatasetIBM Sample Data from the Telco Customer Churn Dataset

Kaggle is the source. It includes account information, service data, and consumer demographics. The target variable is called "Churn."

Using kagglehub, the dataset is programmatically downloaded within the notebook.

Strategy

• Data cleansing and preprocessingScaling features and encoding categorical variables Developing a Classifier for Logistic Regression Assessing model performance through precision
• Interpreting predictions using SHAP values ##Outcomes
• Model accuracy: ~77..64% Contract type, tenure, and monthly charges are the most important factors.

How to Run

1. Clone the repository:

https://github.com/jerem-a11y/customer-churn-prediction