/ml-assignment-3

Primary LanguagePython

ml-assignment-3

Install packages

  1. Install latest version of Anaconda 3 from https://www.anaconda.com/
  2. Clone the project repository from https://github.com/fedme/ml-assignment-3
  3. Open an Anaconda 3 prompt
  4. From the prompt, browse to the folder where you cloned the repository
  5. Run the following command to create the conda environment: 
    conda env create --file=environment.yaml
  6. Activate the newly created environment: 
    activate fmeini3-ml-assignment-3

Alternatively, you can install the latest version of the required packages manually from PIP: pip install numpy pandas scipy matplotlib seaborn scikit-learn yellowbrick

Run the code

Code is divided into multiple python scripts that loosely map the structure of the assignment:

  • python clustering.py runs KMeans and EM clustering and generates plots for them (inside the plots folder)
  • python dimensionality.py runs PCA, ICA, Randomized Projections, and SVD dimensionality reduction algorithms and generates plots for them (inside the plots folder)
  • python clustering_on_reduced_data.py runs KMeans and EM clustering on the dataset with reduced dimensionality and generates plots for them (inside the plots folder)
  • python neural_networks.py trains neural network classifiers on many versions of the Fashion MNIST dataset (base, reduced, augmented with clusters) and generates plots for them (inside the plots folder)
  • The other scripts are used to transform and save data. It is not necessary to run them, since all the data is already present inside the data folder.