Date/Time Date(s) - 22/10/20243:30 pm - 4:30 pm
Location: MDCL 3020
Date & Time: Tuesday, October 22, 2024, 3.30 – 4.30 p.m.
Speaker: Tianyu Guan (York University)
Title: Functional Data Analysis in Sports Analytics and Kinesiology
Abstract: In this talk, I will explore the application of functional data analysis (FDA) to sports analytics and kinesiology. I will first present methods for developing in-game win probabilities for the National Rugby League. By incorporating real-time features, score differentials, and betting odds, we develop a conditional probability framework for predicting win probabilities throughout the course of a match. The second part of my talk investigates how player acceleration changes with age in soccer. With the availability of tracking data from the 2019 season of the Chinese Super League, we calculate an individual player’s maximum acceleration for each single match. The player’s maximum accelerations are treated as incomplete functional data. We estimate the average maximum acceleration by a local linear smoothing method. Finally, I will discuss a sparse estimation method for historical functional linear models. Using this model, we explore the effect of muscle activation on lip acceleration during speech production, which offers new perspectives in kinesiology through FDA.