Machine Learning in Numpy

This repository contains Python implementations of some common machine learning algorithms. Although there are a bunch of similar repositories in GitHub, many of them are hard to understand and lack of optimization. The purpose of this project is to explore the inner mechanism of these algotithms. Therefore all algorithms are built on pure Numpy.

Getting Started

Implemented:

• Linear Regression
• Logistic Regression
• Ridge Regression
• Lasso Regression
• Bayesian Linear Regression
• Decision Tree Classifier
• K Means
• Gaussian Mixture Model
• K Neighbors Classifier
• Naive Bayes
• Support Vector Classifier

Prerequisites

What you need

numpy
sklearn (for validation)
matplotlib (for visualization)

Installing

$ git clone https://github.com/zhangchicheng/ML_from_scratch
$ cd ML_from_scratch
$ python setup.py install

Running the tests

See notebooks folder