cs-433 / ml-project-1-team_bak

ml-project-1-team_bak created by GitHub Classroom

CS433 - Machine Learning Project 1

Authors (team: Team_BAK)

• Elif Kurtay
• Ernesto Bocini
• Abdullah Aydemir

File structure

• report.pdf
  • the pdf of the project report file including explanations of methods and our results.

Data folder:

• train.csv: train data needs to be uploaded and placed here with this name
• test.csv: test data needs to be uploaded and placed here with this name
• submission.csv: result submission

Script folder:

• cross_validation.py
  • File containing functions for splitting data for cross validation to choose best parameters and to perform the final training to retrieve predictions.
• helpers.py
  • File that contains various helper functions for the project generally including loss, gradient, and accuracy computations.
• implementations.py
  • File containing all 6 implementations of ML functions required for the project.
• pre_process.py
  • File containing functions to load and preprocess the data.
• plot_helpers.py
  • File containing plotting functions that are used in Plots.ipynb.
• Training.ipynb
  • File where the training set is used to find the best hyperparameters using k-fold cross-validation
• Plots.ipynb
  • File where the ploting functions are used to show information about the data and about our models' results.
• run.py
  • Main script - training the best model on the train set using the best hyperparameters and using the test set to make predictions for the submission

How to reproduce our results

We assume that the repository is already downloaded and extracted, that the data is downloaded and extracted in the data folder at the root of the program. We further assume that Anaconda is already installed.

Create the environment

Make sure your environment satisfies the following requirements:

• Python 3.7+
• NumPy module
• matplotlib

Run the code

From the root folder of the project

python run.py