Stock trading strategies using machine learning

I'm not even good at coding and this all looks like a mess to me. I wouldn't be surprised if all of this resulted in catastrophic loss.

This is research for a paper. No one is using it for active trading. If the research is successful then these strategies might be applied in the future. If not they're obviously not going to be used.

Sometimes when people are doing research they need to try new things that might seem silly to find new groundbreaking material.

Best one I've seen so far is based on NLP datasets and Elastic https://www.elastic.co/blog/generating-and-visualizing-alpha...

Why would you not host the code in the Github repository?

In this case, it makes sense to host the code and data in Google Drive where you can easily access it with a Colab instance.