Contents
Contents
Under construction!
1. Exchangeability and De Finetti’s Theorem
1.1 Exchangeability
1.2 De Finetti’s Theorem
2. Bayesian Finite Mixture Model
2.1 Finite Mixture Model
2.2 Bayesian Finite Mixture Model
2.3 Inference of Bayesian Finite Mixture Model
2.4 Model Selection for Finite Mixture Models
3. Dirichlet Process
3.1 Definition
3.2 Properties
3.2.1 Expectations, Variances, and Co-variances
3.2.2 Tail-freeness
3.2.3 Self-similarity & Neutral-to-the-right Property
3.2.4 Conjugacy
3.2.5 Marginal and Conditional Distribution
3.2.6 Discreteness
3.2.7 Support
3.3 Constructions
3.3.1 Construction via a Stochastic Process
3.3.2 Construction through a Distribution Function
3.3.3 Construction through a Gamma Process
3.3.4 Construction through Polya Urn Scheme
3.3.5 Stick-breaking Representation(Sethuraman Representation)
4. Dirichlet Process Mixture Model
4.1 Going Nonparametric
4.2 Stick-breaking Process
4.? Inference of Dirichlet Process Mixture Model
4.?.1 Algorithm 1
4.?.2 Algorithm 2
4.?.3 Algorithm 3
Reference
- Ghosal, Subhashis, and Aad Van der Vaart. Fundamentals of nonparametric Bayesian inference. Vol. 44. Cambridge University Press, 2017.
- K. Ghosh, J & Ramamoorthi, R. Bayesian Nonparametrics. Springer Series in Statistics. 16. 2011.
- Talk slide of YeeWhey Teh at The Machine Learning Summer School 2013 at the Max Planck Institute for Intelligent Systems, Tübingen, Germany http://mlss.tuebingen.mpg.de/2013/2013/slides_teh.pdf
- Tutorial lecture of Michael I. Jordan at NIPS’05 http://faculty.dbmi.pitt.edu/day/Bioinf2132-advanced-Bayes-and-R/Bioinf2132-documents-2017/2017-11-30/nips-tutorial05.pdf
- 이재용 교수님, 디리클레 과정 note, 2019.