Ginger Holt, Ph.D. (Hewlett-Packard, USA): ARMA alphabet soup

Everyone (even from outside the department, faculty, university) is warmly welcome at the talk given by our guest speaker Ginger Holt. Her talk will start at 11am at K14 seminar room, ground floor of Building E, CVUT campus at Karlovo namesti 13, Praha. The seminar will take 60 minutes including discussion. Note that this talk will be immediately preceeded by another talk given by another speaker at 10am, see the separate announcement.

Abstract: A tour of basic time series models is given alongside common applications. This talk enumerates, compares, and contrasts the common variants of ARMA (AutoRegressive Moving Average) models in today’s literature. After the basic properties of ARMA models are reviewed, we tour ARMA variants that describe seasonal features, long memory behavior, multivariate series, and changing variances (stochastic volatility). Common tools, modeling steps, and diagnostics used in time series analysis will be introduced. Applications and examples are given in areas of demand forecasting, economics, finance, and ecology.

Bio: Ginger Holt is a Manager in Market Analysis and Planning, Corporate Strategy for the Hewlett-Packard company. She develops predictive models and performs advanced analytics regarding market sizing for their Enterprise business. She is also an adjunct professor in Predictive Analytics at Northwestern University and in Statistics at San Jose State University. She obtained her Ph.D. in Statistics from Rice University where her research was in multivariate time series analysis. She was an Assistant Professor of Systems and Information Engineering at the University of Virginia from 2005-2009 where she conducted research on multivariate time series methodology development. She has researched prediction methodologies in many different applications including computational finance, environmental networks, cyber security, ecology, marketing, and econometrics.