Statistics Seminar: James Flegal
Speaker: James Flegal
Title: Lugsail lag windows for estimating time-average covariance matrices
Abstract: Lag windows are commonly used in time series, econometrics, steady-state simulation, and Markov chain Monte Carlo to estimate time-average covariance matrices. In the presence of high correlation, estimators of the time-average covariance matrix almost always exhibit significant negative bias, leading to undesirable finite-sample properties. We propose a new family of lag windows specifically designed to improve finite-sample performance by offsetting this negative bias in the opposite direction. Any existing lag window can be adapted into a lugsail equivalent with no additional assumptions. We use these lag windows within spectral variance estimators and demonstrate its advantages in a linear regression model with autocorrelated and heteroskedastic residuals. We further consider weighted batch means estimators since spectral variance estimators are too computationally intensive for large simulation output. We obtain the bias and variance results for these multivariate estimators and significantly weaken the mixing condition on the process. Superior finite-sample properties are illustrated in a vector autoregressive process and a Bayesian logistic regression model.
Bio: James Flegal is an Associate Professor of Statistics at the University of California, Riverside. He received his Ph.D. from the School of Statistics at the University of Minnesota in 2008. His research interests include statistical computing, Markov chain Monte Carlo, Bayesian statistical methods, and Monte Carlo standard errors. Prior to graduate school, James attended Northwestern University where he earned a B.S. in Mechanical Engineering in 1999. His professional career outside of academia includes modeling furnace systems with computational fluid dynamics algorithms and designing welded fabrications for a heavy equipment manufacturer.
All School of Statistics Seminars are open to the public. A virtual social will take place from 10:30-11:30AM CDT, following this seminar. To join, visit here.
NOTE: You will need to use a Chrome or Firefox browser to access this virtual platform. The link may not open if you sign in using a mobile device or tablet. The virtual space will not open until approximately 10:30.