External Resources
Video Resources
3Blue1Brown (Brian Sanderson)’s Neural Network series
Brian Sanderson (better known as 3Blue1Brown) has a fantastic four-video series on Neural Networks that covers the mathematics behind basic NNs, including gradients and backpropogation. The first video in the series by itself offers a really good overview of how Neural Networks function as well, however.
3Blue1Brown (Brian Sanderson)’s Essence of Linear Algebra series
Videos 1-9 and 13 of Brian Sanderson’s Essence of Linear Algebra series provide powerful intuition for understanding matrix transformations on a deeper level, and for interpreting linear regression more comprehensively. Remembering all of these videos’ contents is possibly overkill, but having some level of awareness of the concepts discussed in them can make your life easier.
PDF Resources
Professor Martha White’s Machine Learning Handbook
Link to Martha White’s Machine Learning Handbook
Martha has compiled what is essentially a free textbook on machine learning from when she taught/for when she teaches the course. The textbook opens with a fairly rigorous treatment of probability, then talks about optimization and parameters. It formulates essential ML questions in the form of regression questions, and concludes by talking about neural networks and evaluation methods.
Libraries
Python3/SKLearn
If you’re a Python programmer, SciKit-Learn offers a pretty good number of configurable algorithms and tools for analyzing results with them, as well as descriptions of how the algorithms work/how to use them.
Installation: a Pip package is available. To install, run pip install -U scikit-learn
C#/Accord.NET
If you prefer C#, Accord.NET is a powerful option. It sports a number of configurable algorithms, as well as descriptions of how the algorithms work and how to use them.
Installation: a nuget package is available. To install, run Install-Package Accord -Version 3.8.0
with Package Manager or dotnet add package Accord --version 3.8.0
through the .NET command-line interface