/Kaggle_Titanic

Sample code for "Titanic: Machine Learning from Disaster" contest of Kaggle

Primary LanguagePython

Sample code for the Titanic contest of Kaggle.

Elementary sample code for "Titanic: Machine Learning from Disaster" contest of Kaggle.

Requirements

  • Python 2.7.x
  • Scikit-learn
  • matplotlib

How to use

  1. Clone the repository

     git clone https://github.com/sodesign/Kaggle_Titanic.git

  2. Execute

     cd Kaggle_Titanic
 python output_contest_file.py

  3. Submit to Kaggle

    Check the ./result directory and submit to Kaggle it.

Other information

Accuracy

  • About 74-75%
    • Current top accuracy is over 90%.

Algorithm

  • Random forest

ToDo

  • Improve logic for accuracy

    • selection of elements of feature vector
    • graph analysis by human relation (?)

  • Visualization

  • etc.