Skip to content
/ alfred Public

learning about lstms and reinforcement learning against financial datasets

Notifications You must be signed in to change notification settings

j03m/alfred

Repository files navigation

Alfred is an experimental set of machine learning for markets and portfolio management.

The plan for Alfred is to have 3 parts:

  • A series of technical and fundamental analyst models that work on available data to try and project prices and future fundamentals
  • One or maybe more language model analysts that will function as news and research digesters
  • A reinforcement learning model portfolio manager that will ingest data from the ensemble of analysts and try to predict optimal portfolio makeup for some time period N.

I'm currently in the stage of weighing effectiveness vs size for various technical analysts.

About

learning about lstms and reinforcement learning against financial datasets

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published