Skip to content

Latest commit

 

History

History
47 lines (32 loc) · 2.08 KB

README.md

File metadata and controls

47 lines (32 loc) · 2.08 KB

N|Solid

EPFL Machine Learning - Higgs Challenge 2019

About the Project

Using real CERN partcile accelerator data, we implemented a machine learning model able to detect the collision decay signature of a Higgs particle, in order to recreate the process that led to the discovery of the Higgs Boson.

Code Architecture

In this section we explain how our project folder is organised and where to find the needed files.

Data Folder

Please import the train.csv and the test.csv in this folder.

All the CERN particle accelerator data as well as our predictions are found in the /data folder. It consists of:

  1. submission_ridge_reg.csv: file containing our predictions when the run.py code is executed
  2. train.csv : train set data (imported by the user)
  3. test.csv: test set data (imported by the user)

Scripts Folder

All our project implementation can be found within the /scripts folder.

Python Executables .py

  1. run: gives us our best accuracy score on Aicrowd
  2. implementations: contains all the required functions
  3. costs: loss and accuracy functions
  4. cross_validation: all cross validation methods that we used in our process
  5. data_preprocessing : function that help us preprocess the data

Jupyter Notebooks

  1. project_1: runs all different cross-validations for all methods, to obtain an accuracy for each method
  2. datapreprocessing_plot: creates all graphs used for the preprocessing part of the report

Running the code

In this section we explain how to run the code that provided our best submission on the Aicrowd EPFL Higgs Challenge.

  1. Download our .zip project folder and exctract it
  2. Open your terminal and from the script folder execute the following command
python3 run.py
  1. After execution, the obtained predictions are available in the sample-submission.csv file
  2. Upload this file on the Aicrowd Higgs Challenge