automlconfig in Azure Machine Learning Studio RRS feed

  • Question

  • Hi There, 

    I am trying to do some time series forecasting in Azure Machine Learning Studio.   The data is related to production values of certain products. The production varies widely from month to month, but has some general trending over time .... much like a highly volatile stock price over time. 

    Is there any way to do this kind of forecasting in Azure machine learning studio?   I am fully prepared to write a bit of python inside an Execute Python Script module ... but I don't see that there is any way of using sometime like AutoMLConfig to try and find the best algorithm.

    Am I better off just building something in Azure Machine Learning service?     I was hoping to do something quick and dirty in Studio.

    Thanks for any advice the group an offer.



    Thursday, October 10, 2019 6:26 PM

All replies

  • Hello,

    Automated Machine Learning or automlconfig can be used with Azure Machine Learning workspace. With the new Azure ml portal the visual interface or studio components will be integrated along with the workspace to avoid creating a separate workspace for studio. Automated ML visual interface and VM notebooks are also part of this portal. We would recommend using a time series model with ML workspace as these services are getting integrated into one portal.

    If you are looking to use time series forecasting using ML studio you can try this sample experiment from the gallery.


    Friday, October 11, 2019 6:20 AM