Back

Consulting a Machine Learned Model

Hi all,

A component of my trading system uses a Coputational Network Toolkit model (https://www.cntk.ai/). To make a prediction requires sending it transformations of the current period's close, for each instrument in my universe.

I've thought of two ways I could consult my model, and I wanted to know whether either is possible, given QC's data restrictions. 

1. Host the model (<1MB) and CNTK (290MB) within a QC VPS and communicate with CNTK via its C# wrapper. NB: CNTK doesn't require installation, but needs 64-bit Windows; and,

2. Use Azure (https://azure.microsoft.com/en-us/) to deploy the model via a website (there may be other ways to publish models). 

This second option is my preference as I wouldn't need to update software, etc. Might it be possible to whitelist a custom Azure website? I'm pretty sure I can turn off data persistency in Azure, but I'm not sure (yet) how I could provide QC surety of this.

I appreciate that this should be possible using LEAN, but I'd like to leverage the QC Prime service.

Any thoughts appreciated; Thanks,

Gavin.

NB: I'm not looking to use QC compute resources to build models, I can do that on my local machine. 

Update Backtest








Hi @GavinMillynn! You can make a REST interface to your model and have the algorithm consult it for decisions? Would this suit? (The only limitation is that you couldn't use any QC data for the decision making). If your data period is daily you could use an external source like Yahoo to do your model updates.

The backend is Mono / linux based and we couldn't install it if it requires windows.

0

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.


Thanks @Jared, I just needed to know the limits/parameters of the QC system, before I headed down wrong path.

Have a good one mate.

G

0

Update Backtest





0

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.


Loading...

This discussion is closed