search

jankrepl / deepdow

Portfolio optimization with deep learning.
https://deepdow.readthedocs.io
Apache License 2.0
875 stars 136 forks source link

Custom portfolio benchmark #28

Open jankrepl opened 4 years ago

jankrepl commented 4 years ago

It would be nice to have a benchmark that is just some predefined portfolio. One would construct it by passing all the weights.

diegolovison commented 2 years ago

Hi,

Generating synthetic data is harder.

In this case: 1) Could you please upload the daily returns or give me 100 stocks that I can download from yahoo finance or can I pick 100 random stocks from sp500_mean.csv ?

2) The issue has the content benchmark that is just some predefined portfolio? Can I upload a benchmark example that will generate weights "from the optional rebalance count"?

jankrepl commented 2 years ago

Hey there!

I think what I meant here was that it would be nice to create a new Benchmark subclass (see deepdow/benchmarks.py) that would accept a fixed w of shape (n_assets, ) representing predefined weights. The __call__ method would just return it independently of x.

The actual choice of w would be completely up to the user. I hope that makes sense:)

diegolovison commented 2 years ago

What I wish to do it is not related to that. Thanks for the clarification.

guillermop98 commented 2 years ago

Hey @jankrepl, have you been able to get more into to this type of predefined portfolio benchmark?