Often I need to have a nonparametric representation for a measure over some non-finite index set. We might want to represent a probability, or mass, or a rate. I might want this representation to be something flexible and low-assumption, like a Gaussian process. If I want a nonparametric representation of functions this is not hard; I can simply use a Gaussian process. What can I use for measures?

If I am working directly with random distributions of mass then I might want other properties like conservation of mass.

## Completely random measures

See Kingman (1967) for the OG introduction. Foti et al. (2013) summarises

A completely random measure (CRM) is a distribution over measures on some measurable space \(\left(\Theta, \mathcal{F}_{\Theta}\right)\), such that the masses \(\Gamma\left(A_{1}\right), \Gamma\left(A_{2}\right), \ldots\) assigned to disjoint subsets \(A_{1}, A_{2}, \cdots \in \mathcal{F}_{\Theta}\) by a random measure \(\Gamma\) are independent. The class of completely random measures contains important distributions such as the beta process, the gamma process, the Poisson process and the stable subordinator.

AFAICT any subordinator will in fact do. A subordinator is an a.s. non-decreasing Lévy process.

TBC

## Random coefficient polynomials

As seen in random spectral measures

## For categorical variables

## Pitman-Yor

## Indian Buffet process

## Beta process

As seen, apparently, in survival analysis (Hjort 1990; Thibaux and Jordan 2007)

## Other measure priors

Various transforms of Gaussian processes seem popular, e.g. squared or exponentiated. These always seem too messy to me.

## References

*Journal of Applied Probability*25 (A): 175–84.

*Stochastic Processes and Their Applications*43 (1): 9–31.

*Russian Mathematical Surveys*59 (1): 65.

*Stochastic Processes and Their Applications*9 (2): 147–54.

*Artificial Intelligence and Statistics*, 20–28.

*Journal of Machine Learning Research*12 (32): 1185–1224.

*Quantitative Methods for Current Environmental Issues*, edited by Clive W. Anderson, Vic Barnett, Philip C. Chatwin, and Abdel H. El-Shaarawi, 37–56. London: Springer.

*The Annals of Statistics*18 (3): 1259–94.

*Annals of Statistics*33 (4): 1771–99.

*Pacific Journal of Mathematics*21 (1): 59–78.

*Bayesian Analysis*14 (4): 1037–73.

*Bernoulli*28 (1): 638–62.

*arXiv:1905.10733 [Cs, Math, Stat]*, May.

*Bernoulli*20 (3): 1260–91.

*Bayesian Nonparametrics*, edited by Nils Lid Hjort, Chris Holmes, Peter Müller, and Stephen G. Walker. Cambridge University Press.

*Stochastic Environmental Research and Risk Assessment*25 (2): 235–51.

*Annals of the Institute of Statistical Mathematics*41 (2): 227–45.

*Journal of Multivariate Analysis*175 (January): 104560.

*Annals of Statistics*32 (6): 2343–60.

*International Conference on Machine Learning*, 1605–13. PMLR.

*Proceedings of the 22nd International Conference on Neural Information Processing Systems*, 1554–62. NIPS’09. Red Hook, NY, USA: Curran Associates Inc.

*Artificial Intelligence and Statistics*, 800–808. PMLR.

*Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics*, 564–71. PMLR.

*Journal of the Royal Statistical Society: Series B (Statistical Methodology)*61 (3): 485–527.

*Biometrika*85 (2): 251–67.

*Practical Nonparametric and Semiparametric Bayesian Statistics*, edited by Dipak Dey, Peter Müller, and Debajyoti Sinha, 227–42. Lecture Notes in Statistics. New York, NY: Springer.

## No comments yet. Why not leave one?