Combining Predictive Analytics and Experimental Design to Optimize Results
Date: 28 May 2020
Time: 10:30 - 11:00 am BST | 11:30 - 12:00 pm CEST
Can you afford to miss insights hiding within information you might already have? Organizations are today collecting more data than ever. And while some companies may be satisfied with a predictive model that “just works,” they could do so much more to improv understanding, optimize parameters and make better predictions by combining predictive modeling with design of experiments (DOE).
DOE is a method that helps define a data collection plan when the data you already have is not appropriate for solving your problem. Existing data may have been collected for a different purpose so it is not directly relevant, and sometimes the process you are investigating is so new there is limited or no prior data available.
You will learn how to:
- Get the most from large and messy data, even with missing information.
- Use modern screening approaches to find the right potential drivers of performance.
- Compare the potential drivers of performance to pick the best ones for further experimentation.
- Select the right types of experiments to optimize results at the lowest cost.
Register now to attend online.
ABOUT THE SPEAKER
Ben Francis is a Systems Engineer for JMP in the UK team. His job is to understand the science and engineering challenges of industrial organisations so they can improve their analytic excellence and drive their process innovation. Before this role, Ben was Senior Statistician in Unilever’s Digital R&D department, focusing on enabling process understanding across the categories of the CPG giant. This was after a time in academia with a main focus on Statistical Genetics and Big Data analytics. Ben has a Masters and Ph.D. in Mathematics and Medical Statistics respectively.