Learning Bayesian Statistics #2 When should you use Bayesian tools, and Bayes in sports analytics, with Chris Fonnesbeck
5 snips
Oct 23, 2019 Chris Fonnesbeck, senior quantitative analyst for the New York Yankees and associate professor at Vanderbilt, dives deep into the world of Bayesian methods. He illustrates when to effectively employ these techniques and the challenges of teaching them. Fonnesbeck highlights their application in sports analytics, particularly in baseball, alongside marine biology findings. He discusses the importance of skills like programming and understanding priors, while also addressing issues like missing data, showcasing Bayesian's growing relevance across disciplines.
AI Snips
Chapters
Transcript
Episode notes
Bayesian Focus: Estimation
- Bayesian models estimate parameters and quantify uncertainty, rather than testing hypotheses.
- Bayesian outputs, framed as probabilities, enhance communication with non-statisticians.
Bayesian vs. Frequentist
- Bayesian inference directly answers the question of an unknown's value given observed data.
- Frequentist methods, conversely, condition on the unknown, reversing the inquiry.
Building Bayesian Models
- Build Bayesian models iteratively, starting simple and increasing complexity as needed.
- Utilize prior predictive checks to assess model behavior without data.