Yesterday, Microsoft announced Microsoft Azure Machine Learning, previously known as project Passau. Your humble correspondent has been participating in the Preview Program. Basically, Microsoft Azure Machine Learning is a service for self-service cloud-based predictive analytics. You upload your data to the cloud, define datasets and workflows to create “experiments”. Previously, you could create data mining models using the SQL Server data mining capabilities. Microsoft Azure Machine Learning to organizational DM models is what Power Pivot to Analysis Services is. Besides its cloud-based nature, Microsoft Azure Machine Learning offers:
- Data sources – Allow business users to upload source data as files.
- Workflows – Business users can drag and drop tasks to create workflows, such as to perform basic data transformation tasks, remove outliers, train, and score mining models. Users will familiar with SSIS will undoubtedly find workflows similar. Users familiar with SAS data mining will do the same.
- Scalability – You can use Big Data coming from Azure HDInsight and scale out accordingly.
- Endpoints – You can easily publish the predictive results as an ODATA service.
- Algorithms – There are more algorithms than the nine algorithms in SQL Server