In my previous post I outlined a path to learn machine learning that was too idealistic. Being very math heavy, it focused mainly on the goal, not on the steps you take to achieve it. As my main motivation to learn machine learning was being really interested in it, not because I like math, that learning path was not suitable for me and maybe for some of the readers of that post. In this blog, I’ll outline a better method that I use right now.
Basically, this method focuses on learning machine learning top down. Also I’ll be very concise and to the point, as there is already a ton of information online on this way of learning ML.
The first step is to get at least a very basic intuition on machine learning algorithms by completing Andrew Ng’s course. After that, jump directly to the real world with Kaggle. Kaggle is a platform for data scientist who want to compete, learn and socialize. It has a learn subsection, which I recommend for total beginners. After clearing the topics covered in that subsection look into various kernels, starting from the Titanic competition.