The datasets are located in the data/ directory.

File data/craditcard.csv contain the dataset used for this homework.

As stated on the the report, a copy can also be downloaded at: https://www.kaggle.com/dalpozz/creditcardfraud/

File data/train.csv contain the dataset used for this homework.

As stated on the the report, a copy can also be downloaded at: https://www.kaggle.com/c/the-insa-starcraft-2-player-prediction-challenge/data

To get the same result as we do on our report, you must take two steps:

• generate the data
• generate the plots

To generate the data you will need to run the five python scripts corresponding to the five machine learning method studied here.

The five scripts to launch are:

• src/tree.py
• src/perceptron.py
• src/boost.py
• src/svm.py
• src/knn.py

Each of the scripts take as argument the data set to use: "creditcard" or "starcraft"

For example you can launch the decision tree classifier on the starcraft dataset with:

python3 -m src.tree starcraft

You can also get all the .pikle and csv files already generated at: https://github.com/anliec/ML_homework_supervised/tree/master/stats

To generate the plot given in the report you will have to run the following jupyter notebooks:

• plot_tree.ipynb
• plot_perceptron.ipynb
• plot_boost.ipynb
• plot_svm.ipynb
• plot_knn.ipynb

Each one generate the plot for the given machine learning method from the associated .pickle files in the stats directory and try to write the generated plot into the graphs directory.

• Python 3 (tested with Python 3.6)
• Keras
• Sklearn
• Pandas
• Matplotlib
• Seaborn
• Numpy
• jupyter

If you do not already have them installed, it can be done quickly using pip:

pip3 install keras sklearn pandas matplotlib seaborn numpy jupyter