Successfully developed a customer churn classification system using advanced machine learning algorithms in R. This project aims to identify customers who are likely to leave a service based on their demographic, behavioral, and transactional data. The solution includes end-to-end data processing, model building, and evaluation pipelines optimized for real-world deployment.
- Exploratory Data Analysis (EDA): In-depth data visualization using
ggplot2,plotly, and correlation heatmaps. - Feature Engineering:
- One-hot encoding of categorical variables.
- Outlier detection and imputation.
- Class imbalance handled using SMOTE.
- Feature Scaling: Min-max normalization for continuous features.
- Model Building: Implemented and tuned 20+ ML models, including:
- Logistic Regression (Binary & Multinomial)
- Decision Trees, Random Forest, C5.0
- Gradient Boosting (GBM), XGBoost
- Bagging & AdaBoost
- SVM (Radial Kernel)
- K-Nearest Neighbors (KNN)
- Naive Bayes
- Neural Networks (nnet & Keras)
- Regularization methods (Lasso & Ridge)
- Hyperparameter Tuning: Automated model optimization using
caretwith grid search and cross-validation. - Performance Metrics: Accuracy, Precision, Recall, F1-Score, AUC-ROC, and confusion matrices.
- Model Comparison: Automatic evaluation and ranking of models based on performance metrics.
- Deployment Ready: Best-performing model is saved using
saveRDS()for future use.
- Achieved 100% accuracy with KNN and 90%+ accuracy with XGBoost after hyperparameter tuning.
- Robust pipeline enabling easy retraining and scalability.
- Language: R
- Libraries:
caret,ggplot2,plotly,xgboost,lightgbm,nnet,keras,e1071,rpart,randomForest,adabag,MASS,glmnet,smotefamily.
- Telecom, Banking, and SaaS industries to identify customers likely to churn.
- Actionable insights for customer retention strategies.
- Integration with Shiny dashboard for interactive churn visualization.
- Model deployment via REST API or containerization with Docker.