Tracking_with_MLflow

Customer Churn Prediction Tracking with MLflow

This project implements churn prediction using 7 different machine learning models. For each model, we will track various parameters, metrics, and artifacts following MLOps best practices. The project involves:

Features

1. Experiment Tracking with MLflow:

   • Track experiments, parameters, metrics, and model performance.

2. Model Metadata Storage in PostgreSQL:

   • Store model metadata, including hyperparameters and training details, in a PostgreSQL database.

3. Artifact Storage in AWS S3:

   • Store model artifacts (e.g., trained models, datasets) in AWS S3 for scalability and accessibility.

4. Containerized Deployment using Docker:

   • Containerize the model deployment environment using Docker for portability and consistency.

5. Comparison of Multiple Models' Performance:

   • Evaluate and compare the performance of 7 different machine learning models to select the best one for churn prediction.