Navigation

    Quantiacs Community

    • Register
    • Login
    • Search
    • Categories
    • News
    • Recent
    • Tags
    • Popular
    • Users
    • Groups
    1. Home
    2. omohyoid
    O
    • Profile
    • Following 0
    • Followers 0
    • Topics 13
    • Posts 27
    • Best 1
    • Groups 0
    • Blog

    omohyoid

    @omohyoid

    1
    Reputation
    4
    Profile views
    27
    Posts
    0
    Followers
    0
    Following
    Joined Last Online

    omohyoid Follow

    Best posts made by omohyoid

    • I cannot install pandas 1.2.5

      I'm using kaggle to test my strategies on the cloud.
      Today an error occurs:
      KeyError: "cannot represent labeled-based slice indexer for coordinate 'time' with a slice over integer positions; the index is unsorted or non-unique"
      螢幕擷取畫面 2024-06-18 001539.png 螢幕擷取畫面 2024-06-18 001518.png
      Then I check the installation of the packages. I found that pandas 1.2.5 was not successfully installed, which cause the error.
      螢幕擷取畫面 2024-06-18 001720.png
      螢幕擷取畫面 2024-06-18 001752.png
      The error results from the wrong version of pandas, but I cannot install pandas 1.2.5, so I cannot fix the error.
      Does anyone know how to install pandas 1.2.5 on kaggle?
      I may need some help from @support
      Thx very much

      posted in Support
      O
      omohyoid

    Latest posts made by omohyoid

    • RE: ERROR! The max exposure is too high

      @vyacheslav_b Thanks for ur help!

      1. I checked my code, and I found that I loaded the data by using load_spx_data() instead of load_ndx_data()
        But I added the next trading date to the original data in order to write the latest weight. Does this operation cause the warning?
      2. I checked the dataframe of the final weights, and I use the MAX() and the MIN() in excel to check the maximum and the minimum value of the weights, the maximum weight is 0.05, and the minimum weight is -0.05, which is different from the value 0.05882353 showing on the screen.
        螢幕擷取畫面 2025-06-05 001256.png
      posted in Support
      O
      omohyoid
    • ERROR! The max exposure is too high

      @support
      After I tested my strategy,
      I got the warnings below:
      螢幕擷取畫面 2025-06-01 234009.png
      It stated that:

      1. WARNING! The kind of the data and the output are different.
        The kind of the data is None and the kind of the output is stocks_s&p500
        The output will be cleaned with the data kind.
      2. ERROR! The max exposure is too high.
        Max exposure: [0.05555556 0.05555556 0.05555556 ... 0.05882353 0.05882353 0.05882353] Hard limit: 0.1
        Use qnt.output.cut_big_positions() or normalize_by_max_exposure() to fix.

      Do I need to fix the exposure? Exposure of 0.0588 is below the hard limit 0.1, so it seems that I don't need to decrease the current weights. Am I correct?

      BTW, can I just ignore the warning " WARNING! The kind of the data and the output are different." ?
      I don't know what should I fix according to this warning.

      posted in Support
      O
      omohyoid
    • RE: toolbox not working in colab

      @alexeigor I got the same error on kaggle too
      螢幕擷取畫面 2025-04-18 232604.png

      posted in Support
      O
      omohyoid
    • RE: Can I use astronomical data as features for my machine learning model?

      @support Thx for ur reply

      posted in Support
      O
      omohyoid
    • Can I use astronomical data as features for my machine learning model?

      @support I know it sounds strange, but I want to use the distances between the sun and moon and related astronomical data for each trading day as features for my model. The data sources are obtained through mathematical calculations using an open-source Python module, rather than through web scraping.
      Here is the link to this Python open-source module.
      https://rhodesmill.org/skyfield/

      posted in Support
      O
      omohyoid
    • RE: QNT failed to load data after 2006-01-01

      @support I got it. Thanks for ur reply

      posted in Support
      O
      omohyoid
    • QNT failed to load data after 2006-01-01

      I submitted my strategy
      but it failed due to data loading from qnt
      the output shows that the data loading from qnt ranges from 2004-11-01 to 2005-12-30
      the code I submitted is:
      start_date, end_date, prepare_date = '2006-01-01', datetime.now().strftime('%Y-%m-%d'), '2004-11-01'
      data = qndata.stocks_load_spx_data(min_date = prepare_date, max_date = end_date)
      but qnt didn't load the data after 2006-01-01
      @support
      I would like to know how to fix the problem
      螢幕擷取畫面 2024-10-24 115316.png
      螢幕擷取畫面 2024-10-24 120046.png

      posted in Support
      O
      omohyoid
    • Can't download assets

      After I upgrade to the latest version of qnt library(0.0.406)
      An error appears
      "Exception: can't download assets"
      @support Is there a bug in the latest version of qnt?
      ![alt text](螢幕擷取畫面 2024-10-02 130358.png image url)

      posted in Support
      O
      omohyoid
    • RE: How to turn off "WARNING: some dates are missed in the portfolio_history"

      @support Thanks for ur help
      Is there a way to turn off the warning from the outside of the qnt module? I hope that I can just add a line in my strategy code to call something in the qnt module to remove the warning.

      posted in Support
      O
      omohyoid
    • How to turn off "WARNING: some dates are missed in the portfolio_history"

      I'm using partial data for testing the performance for machine learning models.
      qnt module outputs such warning:
      WARNING: some dates are missed in the portfolio_history
      I'd like to turn off this warning
      But I don't know how to do
      Can @support give an example for how to turn off this specific warning?
      螢幕擷取畫面 2024-09-08 225205.png

      posted in Support
      O
      omohyoid
    • Documentation
    • About
    • Career
    • My account
    • Privacy policy
    • Terms and Conditions
    • Cookies policy
    Home
    Copyright © 2014 - 2021 Quantiacs LLC.
    Powered by NodeBB | Contributors