Author here. Yes, it took me well over an year to understand this. All books I read were all theory theory theory. I love maths, and I like reading, so I read t

Author here. Yes, it took me well over an year to understand this. All books I read were all theory theory theory. I love maths, and I like reading, so I read them all! But that went nowhere as to bolster my understanding of how to do ML, after all, if I can't write a ML algorithm then what is the point in knowing the sum of squared errors? I find this approach a bit satisfying, first reading the scikit-learn docs, they are incredible and then going ahead with the theory as you require it, not downplaying the role of theory because there are a lot of parameters to machine learning algorithms. I was hoping that this would help anyone who is new to ML and is struggling to understand _what_ to study and how to study it. Funny thing, I got downvoted on Reddit + a comment saying "seems like you studied ML didn't understand it and gave up, this sends a bad message to your audience"