Machine Learning - Project 1, EPFL

Group:

Rahul Rajesh, Zhuang Xinjie, Chitrangna Bhatt

Background:

https://higgsml.lal.in2p3.fr/files/2014/04/documentation_v1.8.pdf

Objective:

Estimate the likelihood that a given event's signature was the result of a Higgs boson.

Files Attached:

1. linear_models (directory) : this folder contains all the implementation details for the models used. It is written in an object-oriented fashion for ease of use.

   • linear_reg_gd.py : implementation of linear regression using gradient descent
   • linear_reg_lsq.py : implementation of linear regression using normal equation
   • logistic_reg.py : implementation of logistic regression using gradient equation
   • ridge_reg_lsq.py : implementation of ridge regression using normal equation

2. preprocess (directory) : this folder contains all the implementation details for the various preprocessing teachniques used.

   • imputer.py : class that handles imputation for missing or undefined values
   • scaler.py : class that handles normalization of feature matrix

3. implementations.py: method implementations required for the project. Note this file relises on the linear_models directory

4. proj1_helpers.py: various helper methods for the project

5. project1.ipynb: jupyter notebook showcasing the various steps carried out to solve this problem

6. run.py: generates csv file for test set - used for submission to platform

7. report.pdf: final report for project

How to reproduce predictions:

Requirements: Python3, Numpy, Matplotlib

1. Specify input path for data-files in run.py
2. Specify output path for prediction in run.py
3. In your terminal, run python run.py