Are you supposed to implement as many strategies as you can?

Hello, I am writing my thesis on algorithmic trading and have a few questions. Is the name of the game to basically try and test as many strategies as you can? I'm supposed to have a focus on Deeplearning/Machine learning but most strategies dont use any of that.

How can I apply deep learning effectively?

Also I would focus on the 15 min – 4 day timeframe because High frequency trading is to advanced for me and long term trading has way lower sample sizes, thoughts?