📗 Table of Contents

• 📖 Machine Learning Insights into Titanic Survival

• This project explores survival patterns in the Titanic dataset, revealing key insights:

      - Gender Disparity: Women had a 74.20% survival rate, compared to 18.89% for men, reflecting the "women and children first" protocol.
      - Age Impact: Children (under 13) survived at 57.97%, while teenagers had a 41.05% rate.
      - Model Performance: The ensemble model achieved a validation accuracy of 84.17% and a training accuracy of 90.35%, with a Kaggle score of 0.77990.
  

• 🛠 PYTHON

  • Python: Core programming language.
  • Libraries: numpy, pandas, matplotlib, seaborn, scikit-learn.
  • Jupyter Notebook: Analysis in TitanicSurvivalPrediction.ipynb

• Key Features

  • The stacking ensemble will use the RandomForestClassifier, GradientBoostingClassifier, and SVC, using LogisticRegression as the meta-classifier.
  • Feature engineering: Added FamilySize, IsAlone, Title, and binned Age and Fare.
  • Hyperparameter tuning with RandomizedSearchCV and StratifiedKFold cross-validation.

• 💻 Getting Started

  • To get a local copy up and running, follow these steps.
  • Prerequisites
    • Python 3.x
    • Jupyter Notebook
    • Git
  • Setup
    • Clone this repository: git clone https://github.com/Marlyn-Mayienga/TitanicSurvivalPrediction.git
  • Install
    • Install dependencies: pip install numpy pandas matplotlib seaborn scikit-learn
  • Usage
    • Ensure train.csv and test.csv from the Kaggle Titanic competition are in the project directory.
    • Open TitanicSurvivalPrediction.ipynb in Jupyter Notebook.
    • Run all cells to reproduce the analysis, train the model, and generate submission predictions.

• 👥 Authors

  • 👤 Marlyn Mayienga

• GitHub: @Marlyn_Mayienga

• Twitter: @Merl_Mayienga

• LinkedIn: Marlyn_Mayienga

• 🔭 Future Features

  • Incorporate additional ensemble techniques (e.g., XGBoost).
  • Explore advanced feature interactions for improved accuracy.
  • Add visualizations of survival patterns by class and embarkation port.

  Contributions, issues, and feature requests are welcome!

• 🤝 Contributing

  • Feel free to fork the repository, submit pull requests, or open issues for bugs and feature suggestions. Check the issues page for ongoing tasks

• ⭐️ Show your support

  • Give a ⭐️ if you like this project! Your support motivates further development.

• 🙏 Acknowledgements

  • Kaggle: For providing the Titanic dataset and competition platform.
  • Scikit-learn Team: For robust machine learning tools.
  • References:
    • Zhou, Zhi-Hua. Ensemble Methods: Foundations and Algorithms.
    • Elinder, M., & Erixson, O. (2012). Gender, social norms, and survival in maritime disasters. PNAS.

• ❓ FAQ (OPTIONAL)

• 📝 License

  • This project is MIT licensed.

NOTE: we recommend using the MIT license - you can set it up quickly by using templates available on GitHub. You can also use any other license if you wish.

(back to top)