Project 1: Coronary Heart Disease prediction

Team Members:
Amélie Menoud (330315) Lisa van de Panne (395822) Nathan Kabas Kuroiwa (341429)

Aicrowd Team Name: Team_LAN

Overview

This project focuses on developing and evaluating multiple machine learning models to estimate the likelihood of developing coronary heart disease, given a dataset from the BRFSS surveys. We explore several approaches including logistic regression, support vector machines, and other methods, with detailed hyperparameter tuning and model evaluation techniques.

Repository Contents

In this repository, you will find the following files (📌 indicates mandatory deliverables):

• 📌 README.md: This readme file
• 📌 implementations.py: Contains the required functions for Step 2 of the assignment, as well as additional machine learning methods used in other models.
• implementations_notebook.ipynb: A notebook that demonstrates the 6 methods of Step 2 applied to the dataset.
• log_reg.ipynb: A notebook with a detailed implementation of regularized logistic regression, including hyperparameter tuning using K-fold cross-validation and grid search.
• SVM.ipynb: A notebook detailing our Support Vector Machine model, with hyperparameter tuning using K-fold cross-validation and grid search.
• 📌 run.py: Contains the code to reproduce our best submission on AICrowd.
• utilities: A folder with additional code files used to train our models:
  • Data_preprocessing_global.py: Includes tools for data preprocessing, such as data transformation for specific features, standardization, and imputation.
  • helpers.py: Tools to load data, create AICrowd submissions, and evaluate model performance.
  • Hyperparameters_Logreg.py and Hyperparameters_SVM.py: Functions for running cross-validation and selecting optimal hyperparameters.

Finally, you can also find in this repository:

• SVM_submission_file.csv our official test result on aicrowd (not exactly our best one, due to using stochastic methods, but the one with the results included in the report). It is generated by running run.py
• data_with_headers.csv a subsample of the training set, after the pre-processing, for verification purposes. It is generated when running the pre-processing function, as in SVM.ipynb for example.

Remarks

In the notebooks, when loading the data, the path to the folder containing such data must be re-specified by the user, according to your where it is stored locally.