Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
How to automatically find good algorithms with Quantopian framework?

I've been playing with Quantopian for a few days now and it seems like a great framework to write an algorithm, backtest it against some data and see how it performs quickly. One problem for me is that the whole process is quite manual (i.e. you need to enter the code, hit run, see the results, tweak the code and iterate).

Ideally, I want to write a number of different algorithms with a number of different parameters in each. Then, I want to run all different algorithms with different parameter values in parallel (or even in sequence if parallel is not possible) and see which ones perform better than others.

I want to do all of this automatically via code, rather than manually trying out each algorithm and parameter combination myself in the UI. I can't see how this is possible with the current framework. Am I missing something?

5 responses

Hi Mete,
I'm thrilled you've been enjoying the platform. You're right that at this point, it's a manual process to test different parameters in your algorithms. While I don't have a solution for you today, we've recently announced the beta for our brand new Research platform which is where we plan to enable this kind of algorithm and parameter optimization.

I'm hoping to have an example notebook that we can share in the near future. When that is ready, I'll make sure to post it here.

Disclaimer

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 Quantopian. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian 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.

Karen, thanks for the info. Until the research platform is live, I assume I should be able to use zipline directly to automate the algorithm selection process? But probably no, because zipline won't have the data provided by Quantopian.

Beware of over-fitting your data. If you are going to optimize, you should be sure to use out-of-sample data for the optimization process.

Yes, of course. So, are there people using zipline offline for algorithm optimization? If so, how do you deal with the fact that there is no Quantopian data in zipline? What are some good datasets available for zipline?

You might try posting to https://groups.google.com/forum/#!forum/zipline.

https://github.com/quantopian/zipline
Discussion and Help
Discussion of the project is held at the Google Group, [email protected], https://groups.google.com/forum/#!forum/zipline.