Customer Churn Modelling

Applied ML case study on churn prediction, imbalance handling, and model evaluation.

Built a churn prediction model using stratified 5-fold CV, SMOTE, and Gradient Boosting. Achieved 93.2% accuracy and F1 = 0.926, improving baseline accuracy by +8.2 percentage points.

Key Outcomes

93.2% accuracy

F1 = 0.926

Improved baseline accuracy by 8.2 percentage points

View code

Context

Problem and Context

Built a churn prediction model using stratified 5-fold CV, SMOTE, and Gradient Boosting. Achieved 93.2% accuracy and F1 = 0.926, improving baseline accuracy by +8.2 percentage points.

Approach

Approach and Architecture

Churn model using SMOTE + gradient boosting with strong F1 performance.

Implementation

Implementation Details

scikit-learn

SMOTE

Gradient Boosting

Classification

Stratified CV

Results

Results and Tradeoffs

This project is presented as a concise technical overview rather than a full-length narrative case study.

93.2% accuracy

F1 = 0.926

Improved baseline accuracy by 8.2 percentage points

Explore More

Related Projects

Browse adjacent work from the same archive group or jump back to the project archive.

Back to archive

NLP

Multilingual Sentiment & Keyword Analyzer

NLP coursework project combining multilingual transformers, sentiment analysis, and keyword extraction.

Open case study

Computer Vision

Swimming Pool Detection

Computer vision coursework project focused on transfer learning and aerial object detection.

Open case study

Coursework

Used Vehicle Pricing & Valuation Model

Regression case study using feature engineering and tree ensembles on cross-country car listings.

Open case study