Fundamental Machine Learning Algorithms : Supervised Learning

A look at fundamental Machine Learning algorithms implemented in Python3 and C++ at 3 levels:

• _simple.py with as few imports, abstractions, and mutations as possible. No classes. Reinventing the wheel... to a degree.
• _oop.py OOP approach, more common libraries used.
• _sklearn/pytorch.py using Frameworks and modern libraries.

Goal

We need to walk before we run. Nowadays a lot of ML Engineers via YT, bootcamps or other mediums, learn how to use the 'latest buzz word' Framework/API in ML, particularly in NLP. However, as always, there is a huge disconnect between tech recruiters, senior devs, and upper management when it comes to finding a good candidate. Jumping from trendy library to another creates a new level of developers that lack deep knowledge.

I aim to demonstarte my knowledge behind these fundamental algorithms from scratch. From there, we can abstract it away to any level via any library or framework, down to 5 lines of code. But knowing whats under the hood, and able to hack away to problem solve, as plug-and-play doesn't always work for ever situation, is the corner-stone of a valuable developer.

Notebooks

Jupyter notebooks are provided to break down code into smaller parts with explanations.

Algorithms

Python3

• KNN

• Naive Bayes

• SVM (In-Progress)

  C++

• KNN (In-Progress)

• Naive Bayes