Business metrics do a fine job summarizing the past. But to predict how customers will respond in the future, you need predictive analytics. This session will present real-world problems solved with predictive analytics and look at how a predictive model works and how to create one using machine learning algorithms.
Roger Barga is a Group Program Manager in the Microsoft Business Platform Division, where his team builds advanced analytics and machine learning for enterprise customers. In 1997, he joined the Microsoft Research Database Group, working in systems research and product development for database systems, application recovery, workflow, and stream processing.
Roger was a Principal Architect in External Research, responsible for tools and services for data-intensive research, such as the Trident Scientific Workflow Workbench, The Research Information Centre VRE, and Dryad/DryadLINQ on HPCS. He also served as architect and group lead in the eXtreme Computing Group, where his team built services on Azure for technical computing and data-intensive research, such as NCBI BLAST on Azure, the Daytona iterative MapReduce runtime for Windows Azure, and Excel DataScope.