Team Member: Robin Jaccard, Mathieu Despond, Eva Luvison
The Higgs Boson Machine Learning Challenge is a classification challenge where the goal is to predict whether a particle is a Higgs Boson or not. In order to solve this problem, we are using real data from the CERN particle accelerator. We implemented a machine learning model to solve this challenge.
- Python
- Numpy
- Jupyter Notebook
- Matplotlib
- Seaborn
The data must be stored in a "data" folder at the root (same level as run.py). Those files must be stored as zip, exactly as in the repository.
To create a submission file for aicrowd.com, you must go in the scripts directory and run python3 run.py
After running the command a file called result.csv will be created, and can directly be uploaded to aicrowd.com to make a submission.
The different files are :
ExploratoryDataAnalyis.ipynb: allows us to visualized the datas and features distribution to then choose which feature to remove and where were the outliers to handle.helpers.py: contains the functions used load the datasets, give labels and predictions, compute accuracy, create final csv.processing.py: contains the functions used to clean the dataset and preprocess it: split it in 6, delete the NaN values, apply transformations, poly-expansion, add bias and hadle outliers.implementations.py: contains the implementation of the 6 machine learning models asked for this project.GridSearch.ipynb: implements the different version of gridsearch for each method.cross_validation.py: contains the functions to perform a cross-validation.run.py: is used to get the .csv submission file, using the best hyper-parameters found for the chosen function.