Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/cyrof/202-assignment-1


https://github.com/cyrof/202-assignment-1

Last synced: about 2 months ago
JSON representation

Awesome Lists containing this project

README 

        
# ICT202 Assignment 1: Explainable Machine Learning



## Running



First, ensure that you have the following requirements:



- Python `>= 3.11`

- Pipenv `>= 2023.2.0`



Then, run the following in your shell in the repository's root directory (i.e., the directory containing `Pipfile`):



```

$ pipenv install

$ pipenv run jupyter-lab

```



The above should launch JupyterLab's web interface in your default browser.

You can now explore and run the Jupyter notebooks under the `jupyter-files/` directory.



## Project Structure



`Pipfile` and `Pipfile.lock` are specific to Pipenv.

In particular, `Pipfile` specifies this project's dependencies

and is used to create a reproducible environment for better collaboration.



All Jupyter notebooks are found in the `jupyter-files/` directory:

  - `exploration.ipynb`: Contains visualisations and general statistics about the dataset.

  - `preprocessing.ipynb`: Preprocessing of the dataset (train-test split, one-hot encoding, label encoding, standardisation, etc.).

  - `model-selection.ipynb`: Nested cross-validation to select the models and tune their hyperparameters.

  - `training.ipynb`: Training of the models.

  - `lime.ipynb`: Applying LIME to the models.

  - `shap.ipynb`: Applying SHAP to the models.



The `dataset/` folder contains the original dataset in CSV format (`obesityData.csv`),

in addition to its preprocessed variants created and exported by `preprocessing.ipynb`.



The `encoders/` folder contains encoders (one-hot encoder, label encoder, standard scaler) pickled and exported by `preprocessing.ipynb`.

These encoders are used by some of the other notebooks.



Similarly, the `models/` folder contains trained models (random forest classifiers) pickled and exported by `training.ipynb`.



`flake.nix`, `flake.lock` and the `nix/` folder contain [Nix](https://github.com/NixOS/nix)-specific code for a local development shell.



## Github Branch Usage



Please adhere to the following guidelines when using GitHub:

- **Do not edit or push anything directly to the `master` branch.**

- For any features to add, create a feature branch based off the `dev` branch.

- Work on your feature in the respective feature branch.

- Once your feature is complete, submit a pull request from your feature branch to the `dev` branch for review and integration.

- Before submitting a pull request, please ensure your feature branch is up-to-date with the latest changes in the `dev` branch. You can do this by rebasing your feature branch onto the `dev` branch:

    ```bash

    $ git checkout dev

    $ git pull origin dev

    $ git checkout feature/your-feature-name

    $ git rebase dev

    ```

- Resolve any conflicts if they arise during the rebase process.

- Once the rebase is complete and there are no conflicts, push your changes to your feature branch:

    ```bash

    $ git push origin feature/your-feature-name

    ```

- Then, go to GitHub and create a pull request from your feature branch to the `dev` branch.




Thank you for following these guidelines to maintain a clean and organised devleopment workflow.