Reinforcement Learning on a Supervised Dataset

Question

I am trying to do a university project, and my problem is that my professor asked me to do an advanced ML model like RL on a basic supervised dataset, And I tried to convince him about how RL is not right for this too. Is there a way to do RL on a time-based dataset? Or what model (advanced model) do you suggest?

Answer 1

There seems to be quite a misconception at hand about what supervised ML and reinforcement learning is... Where supervised learning is about algorithms that learn an input-output function based on predefined examples, reinforcement learning is about algorithms that learn to achieve a certain goal by interacting with an environment. When your problem is about "time-based data" and you need to use a "advanced model" just for the sake of fancyness (however debatable that might be...) you should find plenty of fancy models in the field of supervised learning alone when you do research in the direction of recurrent neural networks or things like transformer models .