How to Productionize the Machine learning model in Azure Databricks RRS feed

  • Question

  • Team,

    I am trying to explore databricks for Machine Learning. Have some question related to end to end flow.

    Lets say I have my data in Azure blob storage/Data Lake etc. I want to use this as source to train the model.

    If I want to run lets say logistic regression on daily basis. I would need to take the test dataset from database/datalake etc.

    How do I configure this. Also do I need to run the trained model on daily basis ? Or once the model is trained, we can run the prediction on test dataset and store it somewhere.

    Basically I want to productionize this use case of logistic regression.



    Wednesday, September 11, 2019 5:36 AM

All replies

  • Hi Akash , 

    We will explore more the on the databricks and let you on that . 

    On training the model question , yes if you train the model with as current / real data possible , your outcome are going to be better .

    Thanks Himanshu

    Thursday, September 12, 2019 4:55 PM
  • We have not heard back from you , can you please let us know if the issue is resolved if yes , can you please share the resolution ? If you are still having issues we will try to help you .

    Thanks Himanshu

    Wednesday, September 18, 2019 7:28 PM
  • This is not yet resolved. Still waiting for reply.

    Basically I want to know how to productionize the ML models using Azure Databricks.

    Friday, October 11, 2019 8:47 AM