search

wassname / rl-portfolio-management

Attempting to replicate "A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem" https://arxiv.org/abs/1706.10059 (and an openai gym environment)
544 stars 179 forks source link

Question about mean market performance #13

Closed LuhuanWu closed 6 years ago

LuhuanWu commented 6 years ago

Hi!

I was looking through your readme file and I was wondering how you calculated the "mean market performance" in the first plot to be compared to the portfolio agent performance?

wassname commented 6 years ago

I assume someone holds an equal amount of each asset e.g. 33% 33% 33% and plot the return on initial investment.

The old code for it is here. It's just taking a (unweighted) mean of the return for all assets.

LuhuanWu commented 6 years ago

@wassname I see. Thank you~

So what you defined seems identical to the UBAH measurement in the original paper, which is: "the Uniform Buy and Hold (UBAH), a portfolio management approach simply equally spreading the total fund into the preselected assets and holding them without making any purchases or selling until the end (Li and Hoi, 2014)"

wassname commented 6 years ago

Yup that sounds right!