
Stephen Smith
Mariner Innovations
The Top Eight Ways We Fail with Predictive Analytics in Business
Presentation Synopsis: Far too often predictive models are never deployed by the business end user they were intended for. And powerful models are misapplied, leading to expensive mistakes. To understand why, we conducted more than thirty in-depth interviews with predictive analytics experts across health, finance and other industries. That research resulted in a compilation of the top eight ways that these practitioners have seen predictive analytics models fail in their businesses. These failures were reviewed and analyzed for commonalities and culminated in best practices in technology, process and organization that these best-in-class practitioners employ to keep predictive analytics models safe and productive.
Speaker’s Bio: Mariner Innovations Inc. would like to present Mr. Stephen J. Smith who is the research director for data science at the Eckerson Group. His unique perspective comes from his real-world experience in leading the teams that built the predictive analytics products Darwin (Thinking Machines), Discovery Server (Dun & Bradstreet), JogNog (G7 Research) and Optas. He has written extensively in the fields of data mining, machine learning, parallel supercomputing, text understanding and simulated evolution and published in the Communications of the ACM and IEEE Transactions on Pattern Analysis and Machine Intelligence. He has also authored two books for McGaw-Hill: Data Warehousing, Data Mining and OLAP and Building Data Mining Application for CRM. He has patents in the fields of predictive analytics, educational technology, big data, and machine learning. He currently serves as a lead mentor at the MIT Venture Mentoring Service and holds a BS in Electrical Engineering from MIT and an MS in Applied Sciences from Harvard University.
Mariner Innovations
The Top Eight Ways We Fail with Predictive Analytics in Business
Presentation Synopsis: Far too often predictive models are never deployed by the business end user they were intended for. And powerful models are misapplied, leading to expensive mistakes. To understand why, we conducted more than thirty in-depth interviews with predictive analytics experts across health, finance and other industries. That research resulted in a compilation of the top eight ways that these practitioners have seen predictive analytics models fail in their businesses. These failures were reviewed and analyzed for commonalities and culminated in best practices in technology, process and organization that these best-in-class practitioners employ to keep predictive analytics models safe and productive.
Speaker’s Bio: Mariner Innovations Inc. would like to present Mr. Stephen J. Smith who is the research director for data science at the Eckerson Group. His unique perspective comes from his real-world experience in leading the teams that built the predictive analytics products Darwin (Thinking Machines), Discovery Server (Dun & Bradstreet), JogNog (G7 Research) and Optas. He has written extensively in the fields of data mining, machine learning, parallel supercomputing, text understanding and simulated evolution and published in the Communications of the ACM and IEEE Transactions on Pattern Analysis and Machine Intelligence. He has also authored two books for McGaw-Hill: Data Warehousing, Data Mining and OLAP and Building Data Mining Application for CRM. He has patents in the fields of predictive analytics, educational technology, big data, and machine learning. He currently serves as a lead mentor at the MIT Venture Mentoring Service and holds a BS in Electrical Engineering from MIT and an MS in Applied Sciences from Harvard University.