I realized a lot of people keep pursuing and learning Competitive Programming because they get addicted to Rankings & Ratings. In this blog i have tried to help you exploit your same existing mechanism and learn Machine Learning in the process while pursuing it a competitive way.
I have drawn examples from around 10 kaggle Competitions and illustrated what works and what doesn't , what is important and what is not and how much time usually it takes for each step while competing. I have also heavily linked through the blog post and it contains more than enough links for you to start competing on Kaggle or Analytics Vidhya seriously.
Link to original Blog Post on "Introduction to Competitive Machine Learning"
Also last event by Threads, Felicity'18 called "Kings of Machine Learning" is still live. There is Rs 15,000 cash prizes and 20 T-shirts up for the Grab. If you enjoyed participating in other threads events like Gordian Knot, Codecraft, Fools Programming then probably you will enjoy this too.
Please keep up the good work. I have been trying to learn ML for years in vain since I find it so different from the fun and exciting CP. I cannot wait to see some Codeforces-like online judge for ML.
Wow! Article is packed with lots of information concisely. Reading even just summary at end alone is useful compilation of insights, thanks for the great post.
It would be great if Codeforces did Machine Learning related rounds. Hopefully the mini marathon will be something like this. It would also be great if for these rounds instead an editorial there would be a tutorial on how to make an AI that works for this specific problem (not perfectly of course) and maybe demonstrate how a certain machine learning algorithm works, when it should be used and how to code one.