Videos

Utilizing Numerical Optimization to Improve the Utility of Data-Based Forecasting Models

Presenter
October 13, 2017
Abstract
The explosion of the availability of data has been both a blessing and a curse. In this talk, we will examine how data can be a powerful tool for driving change if used correctly. We will discuss how the tools of data science can be used to give a statistical basis and robust credibility to a problem that at first glance may seem to be only described by tribal knowledge and how best to utilize data to describe the current environment and lend a historical context to goal setting and forecasting exercises. Attention will be paid to the increased scalability, complexity, and disparateness of the data being generated and analyzed is leading to significant changes in how models are being formulated and utilized. Moreover, we will describe how to incorporate a data-based model into an optimization framework that will illustrate possible outcomes and demonstrate selection of the variables needed to make meaningful change. The techniques described will be demonstrated for the problem of underrepresentation of women in the STEM fields. Bio Dr. Genetha Anne Gray is an analytics research scientist at Intel Corporation where she works on the design and development of machine learning algorithms and analytics capabilities. Previously, she was part of the Talent Intelligence & Analytics organization where she analyzed talent supply chains, studied career progression, and modeled the changing representation of women in the workforce. Before joining Intel in 2014, Genetha spent 12 years as a member of the technical staff at Sandia National Labs in Livermore, CA. There, she has worked on problems related to the electrical and mechanical engineering of systems, the storage of nuclear waste, groundwater remediation, cyber security, and energy including renewables integration and grid operations. She has a Ph.D. in Computational & Applied Mathematics from Rice University and specializes in analytics techniques for decision making under uncertainty including optimization, data fusion techniques, model validation, and uncertainty quantification. She has co-authored more than 25 research publications, given more than 50 presentations at conferences and at universities and is the co-author of a recently released text book on environmental modeling.