A quick question

[DeepAnonymous]

Hi,

I saw you re-publish this repo, is there any different compared to this https://github.com/DVampire/PortfolioManagement

Also, just a quick question, is the idea of Maskable mean that we could input any stocks and it returns a weight vector? Or we still need to train and infer the same stocks with fixed ordering?

Thank you.

[DVampire]

Sorry for the late reply, this repository is the same as https://github.com/DVampire/PortfolioManagement. I'll be deleting https://github.com/DVampire/PortfolioManagement later, so I’m migrating to the new repository EarnMore.

No, this work of ours is semi-open stock pools. For example, training on a set of 30 stocks can be tested on any subset of those 30 stocks. As for the advantages of this you can refer to our article https://arxiv.org/abs/2311.10801. Adding a new stock at will or deleting a stock from the full open stock pool is what we do afterwards.