Kaggle Intro

Installation

1. (windows) install git
2. Install miniconda
3. Install packages
   • conda install numpy scikit-learn pandas matplotlib seaborn
   • conda install jupyter
   • pip install kaggle-cli
4. Install XGBoost
   • Windows
   • Linux/OSX
     1. Install lib

        RUN git clone --recursive https://github.com/dmlc/xgboost && \
        cd xgboost && cp make/minimum.mk ./config.mk; && \
        make -j4 
        

     2. Install python wrapper package

        cd xgboost/python-package; python3 setup.py install && cd ../..
        

5. Register on Kaggle (if not)

Follow the Jupyter notebook

1. Open this repo git clone https://github.com/imsedim/intro.kgl.git

2. Go into the directory cd intro.kgl

3. Download competition files into folder input

   mkdir input
   cd input
   kg download -u <username> -p <password> -c porto-seguro-safe-driver-prediction
   cd ..
   

4. Run jupyter notebook