Sklearn dataset predictions

This repository contains the predictions, and plots for the datasets included in the scikit learn library by default and also some other datasets from kaggle or other sources.

🛠️ Tech stack used

Pandas: For the data manipulation
Matplotlib: Doing plotting
Numpy: As a dependency for Pandas
Scikit-learn: The most important library for ML

❓ How to run this locally

NOTE:

Before cloning this repo, you need to ensure you have GIT LFS installed on your local system. Because this repository contains several *.csv files, which are quite large and aren't accepted by github directly. Sorry for this inconvience.

Steps for running locally:

Run for Testing

As the virtualenv for separating the dependencies, I've gone with pipenv for it. It's really modular and easy to use.

Use pipenv shell to activate the virtualenv and then execute the python commands to run the files and display accuracy.
Run for development and contributing

We also encourage people to support this repository by contributing, and keeping it alive. But note that we follow certain steps to ensure code is clean, organized and readable using linting with flake8. We also encourage using pre-commit for pushing clean code.

Steps to set up:
- Install dependencies: pipenv update -d
- Setup pre commit: pipenv run precommit
- After changes, try linting: pipenv run lint

Datasets implemented

Diabetes:

This dataset consists of 9 columns. The target value which has to be predicted is diabetes This is a classifier problem, where the value of diabetes in boolean, but in integer format.

Algorithm used for the problem: GradientBoostingClassifier

Accuracy achieved: 0.74

🤝 Contributing

Contributions, issues and feature requests are welcome. After cloning & setting up project locally, you can just submit a PR to this repo and it will be deployed once it's accepted. The contributing file can be found here.

⚠️ It’s good to have descriptive commit messages, or PR titles so that other contributors can understand about your commit or the PR Created. Read conventional commits before making the commit message.

And, for contributions we have a Branch named dev, So if you're interested in contributing, Please contribute to that branch instead of the main branch.