Repo where I recreate some popular machine learning models from scratch in Python

A repository where I recreate some popular machine learning models from scratch in Python for educational purposes.
Explore the docs »

View Demo · Report Bug · Request Feature

This repository is for educational purposes. While the efficiency of the code or completeness is not as good as the one you could find in a library like sklearn, the simplification of the algorithm allows you to really understand how the algorithm works, and why it's being used for a particular dataset.

Every subfolder contains a file demo.ipynb, which contains a practical application of the algorithm. It loads a dataset, perform a basic EDA, and fits the model.

Moreover, I published an article explaining in granular details every model present in this repo. I highly recommend reading them, as I go through on how the model works, the math behind the model, its benefits, assumptions, and cons, and most importantly, a breakdown and explanation of the code to build it.

• AdaBoost
• Adam
• CART
• ID3 - Decision Tree
• K-Means Clustering
• KNN
• Naive Bayes Classifier
• Neural Networks
• Fine-Tuning Deep Neural Networks
• Principal Component Analysis
• Random Forest
• Stochastic Gradient Descent
• Support Vector Classfier
• XGBoost

Follow these steps to get a local copy up and running:

1. Clone the repository:

   git clone https://github.com/cristianleoo/models-from-scratch-python.git

2. Install Python packages. You can use pip or conda, depending on your setup:

   pip install -r requirements.txt

3. (Optional) Run the tests to verify that everything is set up correctly:

   python -m unittest discover

Now, you're ready to use the repository!