01 January 1970 0 4K Report

Hi fellow researchers,

I am at the stage of my project where I need to bounce off ideas from someone who shares similar research interests before finalising the dataset and methodology. 

Questions such as:

  • For the simplest kick-start possible, should I employ Elementary Cellular Automation (ECA) or Generative Adversarial Networks (GAN)?
  • Assuming ECA is the simplest starting point, does it make sense to encode stock prices using 8 or 10 bit binary strings as inputs into Elementary Cellular Automata (ECA)?
  • Perhaps, break down a time series into multiple 100 sub-sequences to try to infer overlapping rules that might generate the same output after iterating the ECA 100 times?
  • How do I link ECA patterns to market factors or otherwise since investors dont like black boxes?
  • Please let me know if you need any other details before committing your time in this project.

    Thank you.

    Yeu

    Reference:

    Causal deconvolution by algorithmic generative models | Nature Machine Intelligence

    Similar questions and discussions