Machine Learning Algorithms

Algorithms written for foundations of machine learning class at AMMI

This repo contains Jupyter notebooks implementing various machine learning algorithms you will come across as you start your machine learning journey. Most are implemented as python classes which encapsulate all the various parts of the algorithm.

Getting started

To learn more about the algorithms, you may require to start with a machine learning book. I wrote a Foundations of Machine Learning Note during the course which might be helpful.

To get a copy of the notebooks, run the following command in your terminal;

git clone https://github.com/ogunlao/foundations_of_ml

The code above fetches the foundations_of_ml repository and saves it into foundations_of_ml in your current directory.

Algorithms covered

Advanced Numpy - Just to get comfortable with some numpy concepts
Model Selection Algorithm - Backward Selection
Coordinate descent algorithm
Kullback–Leibler divergence
Linear Algebra - Vectors and their properties
Linear Regression
Perceptron
Binary Logistic Regression
Logistic Regression with L1 Regularization
Logistic Regression with L2 Regularization
Multiclass Logistic Regression - Softmax Regression
Logistic Regression with Newton Algorithm
Gaussian Discriminant Analysis (GDA)
Linear Discriminant Analysis (LDA)
Quadratic Discriminant Analysis (QDA)
Gaussian Mixture Models (GMM)
Bernoulli-NaiveBayes
MultiClass - NaiveBayes
K-Fold Cross Validation
KMeans Clustering
Newtons Method
3D Plots tutorial
Principal Component Analysis

Most algorithms use numpy and scipy internally for vectorized implementations. Note that these algorithms are not optimized for deployment but serve to explain the concepts in a clear way. For deployment, check out the scikit-learn python library for a start.

Licensing

This project is licensed under MIT license.

Enjoy.