Granger Causality and Lasso Penalty: Nerding Out To The Beautiful Game by Onder Kilinc ’23

Granger Causality and Lasso Penalty: Nerding Out To The Beautiful Game by Onder Kilinc ’23, Statistics Colloquium, Wednesday, April 19, 1:10 – 1:50 pm, North Science Building 114, Wachenheim.

Abstract: Granger Causality (also known as G-causality) is a tool to study the usefulness of one variable in predicting another for time series data with applications in biostatistics, physics, and, most commonly econometrics.  An adjacent topic that advances our application of G-causality, Lasso Penalty, helps control for various assumptions by shrinking regression coefficients to 0. This talk will briefly touch on both topics and the breadth of their uses before discussing a unique application to the beautiful game.  G-causality has often been used to predict variability in soccer (football!), this talk will discuss these applications with a particular focus on the limitations of the tools.