Questions? Feedback? powered by Olark live chat software
Hopp over navigasjon

Using Azure Machine Learning with SQL Data Warehouse

Posted on 24 november, 2015

Program Manager, SQL Engineering

Azure Machine Learning allows you to build predictive models using data from your Azure SQL Data Warehouse database and other sources.

Azure Machine Learning is a powerful cloud-based predictive analytics service that makes it possible to quickly create and deploy predictive models as analytics solutions. It allows you to run T-SQL queries on your SQL Data Warehouse database to process, sample and load data into the machine learning environment. You can then create, train and score predictive models on this data by choosing from a large library of ready-to-use algorithms or by writing your own custom R and Python scripts. Once you scored the model, you can write the results to your SQL Data Warehouse database and visualize using Power BI. You can also operationalize the model by publishing it as a web service.

Visit the Machine Learning Studio and click the Get Started button. You can choose either the Guest Access or sign in with your Microsoft account. Once you are connected,

  • Start a new experiment by clicking +NEW and then Blank Experiment
  • Read data from your Azure SQL Data Warehouse using a T-SQL query
  • Create, train and score a predictive machine learning model
  • Write the result to your Azure SQL Data Warehouse, or
  • Publish the model as a web service

Want to try building machine learning models on top of data from your SQL Data Warehouse database? Follow the tutorial below or check out the Analyze with Azure Machine Learning article.

Additionally, the PolyBase feature in SQL Data Warehouse allows you to mash-up relational data in your database with semi-structured data from your Azure storage accounts. You can build machine learning models on these combined data sets in addition to creating external tables in your database that reference data in Azure storage. Follow the tutorial below to get started with PolyBase or check out the Load with PolyBase article.