Samaj

A tool for understanding :)

How to Run Locally

For best results, use Python 3.9. Please use python -m pip install -r requirements.txt to install the dependencies.

What's Included Here

In this repo, I'm implementing classic machine learning algorithms from scratch with NumPy. Here's what we have so far:

1. Exploratory Data Analysis - see the analysis package.
2. Linear/Polynomial Regression via OLS - see the models.supervised.regression.ols and models.supervised.regression.linear_via_backprop modules.
3. Unsupervised Learning via K-Means Clustering - see the models.unsupervised.clustering.kmeans module.
4. For supervised models that use backprop: we have monitor.py, to record the learning curves.
5. Optimizers: so far has gradient_descent.