PCA Approach

[PavelJelen]

Hi,
I was reviewing a code for PCA Approach (pca_approach.py) and I was wondering about the _generate_signals function (line 197).
It says:
":param size: (float) Number of units invested in assets when opening trades. So when opening
a long position, buying (size) units of stock and selling (size) * betas units of other
stocks." (line 217)

However, if you look at line 236-238, I can see that it buys (size) for the most recent value of a specific "ticker" (line 237) and then it goes short (size) on all the values for the same "ticker" (line 238), essentially overwriting the previous line.

My understanding is that it should short all the other stocks (and not the same one) as per documentation.
Would you please shed some light on that?
Thank you

[Jackal08]

Hi @PavelJelen,

Thank you for raising this.

At the moment the focus is to imporve the accessability and get the repo into a strong open-source CI/CD setup.

Once that is done, we can review technical details like this.

Best regards
Jacques

[tinserrano]

Hello!

I was trying to install the library with the official docs both in local virtual environment and in googlecolab. But it doesn't work for me because of different versions of pandas, tensorflow and other problems. I tried to install the versions that the framework requires, but it didn't work. Can you help me or do you have an updated guide to install?

[ben9809]

Hi @tinserrano,
we are releasing a new version that should solve this problem and that is also compatible with Python greater than 3.8.