›

Api calls

dk490971dk490971 Posts: 7
edited August 2018 in Quick Question

I was wondering how within the cloudquant platform would I be able to use predictions from a ML decision tree ensemble (scikit or gxboost etc.)
Via a api call from Google Cloud ML or some other source/ method to trigger my actual Buy and Sell events. Ensembles do not translate into standard buy and sell rules
so I need to rely on a single scalar prediction or prediction probability.

Also, I was wondering of there is a pattern in the strategies that are being funded by cloudquant regarding time frame. Are most based on short term long term signals or scalping / intraday or ?

Thank you in advance

Comments

  • spotlight1spotlight1 Posts: 2

    Best bet for integrating your machine learning predictions is to either save them to a signals file and save that to user data then read in when needed based on symbol and timestamp, or to pickle the actual classifiers then save them into user data and load them during the beginning of your model (sklearn is supported XGBoost should be supported shortly, along with tensorflow and lightgbm). Right now you have to copy and past code into a user data file - which isn't ideal but works for the time being. Happy to share code for either if that isn't clear.

    I'm not the best one to speak to strategies getting accepted, but I'd say holding periods from a few minutes to 10 days is what we're looking for.

Sign In or Register to comment.