CS-433 Machine Learning Project

By Ahmed J., Karim H. and Ayman M.

This code is our solution to EPFL's CS-433 Machine Learning course competition : EPFL Machine Learning Higgs

Results

Our implementation achieves an accuracy of 81.5%.

Files

• proj1_helpers.py: file provided by the teaching team.
• implementations.py: contains our implementations of the labs' functions
• cross_validation.py: contains our cross validation code
• cross_validation_run.py: performs cross validation to produce the best hyper-parameters for our model (stored in best_degrees.npy, best_lambdas.npy)
• data_manipulation.py: contains our helper methods to perform data manipulation and feature engineering
• run.py: contains the code we use to train our model and make our predictions (needs best_degrees.npy, best_lambdas.npy)
• best_degrees.npy: contains the best degrees for feature augmentation. Can be reproduced by running the script cross_validation_run.py
• best_lambdas.npy: contains the best lambda for our machine learning model training. Can be reproduced by running the script cross_validation_run.py
• submission.csv: tabular file containing our predictions. Used for the submission on the challenge website

Execution

To execute our code, the dataset must be downloaded from here. The .csv must be extracted from the archive and placed in a folder called data placed on the root of the project.

Contact

In case any help is needed:

• Ahmed Jellouli : ahmed.jellouli@epfl.ch
• Ayman Mezghani : ayman.mezghani@epfl.ch
• Karim Hadidane : karim.hadidane@epfl.ch