This project predicts Sleep Efficiency (ratio of time asleep to time in bed) using machine learning techniques on lifestyle and sleep-stage data, and deploys the final model as an interactive Shiny App.
-
End-to-End Pipeline
- Data preprocessing (cleaning, imputation, encoding)
- Feature selection (RFE, Lasso, ANOVA, Forward Selection)
- Model training and validation (Decision Tree, Random Forest, SVR, Gradient Boosting)
- Evaluation metrics: R², RMSE, MAE
-
Best Model: Random Forest + RFE-selected ~8 features (R² ≈ 0.90)
-
Deployment: Interactive Shiny app for real-time predictions
-
Explainability: Feature importance & interpretability methods included.
The app allows users to input:
- Light/Deep/REM Sleep %
- Awakenings
- Alcohol consumption
- Age
- Smoking status
- Exercise frequency
Output: Predicted Sleep Efficiency with optional interpretation.
- Lasso Regreesion as Feature Selection Bar Plot for different Models.
- ANOVA as Feature Selection Bar Plot for different Models.
- RFE as Feature Selection Bar Plot for different Models.
- ForwardFS as Feature Selection Bar Plot for different Models.
- Improved imputation (KNN or missForest)
- Hyperparameter tuning (grid or Bayesian)
- Alternative models (XGBoost, LightGBM)
- Model monitoring for drift in production
Open-source under Apache 2.0 License.