/HiggsBosson_ML

Binary Classification using basic ML methods without external libraries

Primary LanguagePython

Machine learning Project 1

  • Wenuka Gunarathne
  • AurĂ©lien Balice-Debbas
  • Giorgio Savini

Structure

We have provided the foloving files and folders:

  • scripts

    • helpers.py : Helper functions for cleaning, finding parameters and utility functions (gradient, mean square error....). Needed for run.py and implementation.py
    • run.py : Containing the executable to produce our best predictions.
    • implementations.py : The implementation of the requested functions.

  • data : ** Not included as the file is large, please add this before running the code **

    • train.csv : CSV containing training data
    • test.csv : CSV containing the test data

  • Project_structure.ipynb

  • report.pdf : The final report

  • README.md. : This readme file

scripts/run.py produce the best prediction in a file called best_predictions.csv

** NB ** : It takes a long time to run.

Requirments

You need to have folowing requirments to run this code smoothly,

  • Python 3.5+
  • Python numpy package
  • Python csv package
  • Python tqdm package (only used for visualization (for progressbar))

Results

You can find the final predictions in the ./data/best_predictions.csv file