Bring your laptop to this session! Azure Machine Learning Studio simplifies machine learning experimentation. In this hands-on tutorial, we will go through the end-to-end process of building, evaluating, fine-tuning and deploying a scalable predictive modelling web service using Azure Machine Learning Studio. By the end of the tutorial, attendees will have a deployed predictive modeling UI of their own, similar to one of these: http://demos.datasciencedojo.com/. All attendees will go through the following hands-on exercises:
• Exploring, visualizing and cleaning a dataset • Building and fine-tuning a predictive model • Evaluating and comparing predictive models • Deploying a predictive model as a full managed, scalable web service • Exposing the deployed predictive model as a web UI.
The Azure Marketplace provides machine learning services we can use with Python Tools for Visual Studio to make better applications. This video shows how to use the Recommendations service to provide better purchasing suggestions to users of an eCommerce site.
The decisions you make during the design phase of an Azure solution tend to stick with you as you gain momentum delivering features. If you have a green field application, you have the luxury of weighing those decisions up front. If you are brown field - understanding the choices may help you lead to a migration path that creates efficiencies for you later on. In this session, Michele will explore decisions related to the various PaaS and IaaS features of the platform. Why would you choose a Web App over Virtual Machines? Do you need API management? When is SQL Server or Azure SQL Database appropriate? Should you use Web Jobs or Worker Roles for async and scheduled tasks? How do you choose between Service Bus queues, topics or event hubs? Is Azure AD the right fit for your identity requirements? All this in just one session - you may need a coffee first!