
PhD Candidate, Sports Data Science
- Incorporating external information
in modelling with BART
- Paddle Australia
- QUT and AIS/QAS PhD Scholarship
- Paul Wu, Kerrie Mengersen
Amir Rezaeian is a mathematician at heart and a data scientist by practice. His main interests (in no specific order) are discrete mathematics, machine learning, statistical modelling, AI, generative AI and generally all interesting and novel ways of making sense of data.
What made you interested in this project and/or program?
Sports play a significant role in the Australian culture. With the advent of new technologies in collecting data, the availability of computing power and the development of new statistical and mathematical methods, we can support our athletes, coaches and sports scientists in ways not possible before. It is particularly interesting for me to see the output of my research can have a tangible effect on the way we analyse sports data and could potentially provide insights not possible without the power of statistical modelling.
What research question(s) are you looking forward to exploring?
In outdoor sports, environmental conditions can have a tremendous effect on performance times. In Paddling sports like kayaking and canoe racing, factors like wind strength and wind direction, temperature and water salinity can impact how fast an athlete finishes the race. In Paddle Australia, one of the questions is how to compare different finishing times and “correct” them for these environmental conditions to have a less biased comparison of the performances. We have employed Bayesian Additive Regression Trees (BART) modelling to study these differences. BART is particularly of interest because of its high accuracy, its ability to provide insights about important covariates and importantly providing a probabilistic paradigm for the predictions which can be useful for decision-makers in sports.
Fun fact about yourself?
Apart from mathematics, data and sports, I am an avid fan of classical Persian literature!