poudel-bibek / k-fold-cross-validation

This is a coding assignment submission for the Data Mining class

Major Code Dependencies:

Scikit-learn
PyTorch
Numpy
Matplotlib
Scipy

Jupyter Notebook

How to run the code:

If you are unable to install the dependencies on your system, 100% of the code is contained in 4 jupyter notebooks so, you can render them in this GitHub repository: https://github.com/Bibek-Poudel/Coding_Assignment_C3

Else,

• Step 1: Install Jupyter Notebook and other dependencies in your system.
• Step 2: You will find 4 Notebooks, each one concerning with a different task in the assignment:
  • Notebook 1: Cross Validation for Binary Classification (performs 5 fold cross validation on all binary classification datasets given)
  • Notebook 2: Binary Classification (performs binary classification on all datasets using the best found values from Notebook 1)
  • Notebook 3: Cross Validation for Multi-Class Classification (performs 5 fold cross validation on the multiclass classification dataset)
  • Notebook 4: Multi Class Classification (performs multi class classification on the best found values from Notebook 3)

Datasets

The dataset files are included in GitHub but dropbox submissions max out at 48 MB. So, to run the jupyter notebook locally in your system, inside the datasets folders, place the uncompressed versions of dataset files (.npz or .mat files) in the respective subfolders given below:

folders:
datasets\bi-class\
datasets\multi-class\