This repository implements a couple of machine learning algorithms from scratch. The idea is simple; by implementing them from scratch, we might understand a little better what they actually do and how one can interpret the different algorithms. This intends to only implement the core functionality features and omits a lot of the advanced model-specific features you find in libraries like e.g. scikit-learn.
- KNN
- Linear Regression
- Logistic Regression
- Naive Bayes
- Perceptron
Clone this repository and then setup locally with pip from within the ml_models_from_scratch directory
cd ml_models_from_scratch/
python3 -m pip install .Each ML algorithm comes with a corresponding Jupyter notebook that is available in notebooks/ for you to try out.