Recreating machine learning algorithms in code from scratch
Why write Machine Learning algorithms from scratch when there are modules like scikit-learn that can do it more efficiently?
- The main purpose of writing Machine Learning algorithms from scratch is to give you a basic understanding of the inner workings of those algorithms.
- If you want to push the limits on performance and efficiency of a specific Machine Learning algorithm you need to know how it works in the first place, which is why coding it from ground up is cruical for any Machine Learning Engineer/ Data Scientist.
Some basics in Linear Algebra is required to understand the math involved.(It's pretty simple though)
Some of the packages required to run the code are the following: numpy, matplotlib, pandas, sklearn and their dependencies.
The code written in this repository are the result of taking a Machine Learning course at USC and also few tutorials online.