Welcome to MathOverflow

Question

1

Hi, i'm looking to get into nonparametric bayesian techniques but I'm having problem understanding what's going on in the definition of the Dirichlet process or how it works. So what does P ~ DP(α*P₀) mean?

What does a distribution P looks like? Is the samples being used, X_i ~ P?

flag	asked Dec 16 2009 at 21:03 mathfool 13●2

2

It might help if you let us know what definition you have in front of you. – Michael Lugo Dec 16 2009 at 21:08

Andrew · Accepted Answer · 2009-12-25 06:09:57Z UTC

0

I'm a fan of Yee Whye Teh's tutorials, listed under "Short Courses" here: http://www.gatsby.ucl.ac.uk/~ywteh/teaching/teaching.html

You can also watch the video on videolectures if you want an explanation to accompany the slides.

answered Dec 25 2009 at 6:09

Andrew
211●1●7

Piotr Miłoś · Answer 2 · 2009-12-16 21:35:45Z UTC

1

Are you taking about the Dirichlet-Poisson process? If yes then a nice introduction is in http://arxiv.org/abs/0909.3985 .

It is too long to put here (you have to read first about random partitions) but it gives some intuitions.

answered Dec 16 2009 at 21:35

Piotr Miłoś
746●3●9

mathfool · Answer 3 · 2009-12-17 06:32:22Z UTC

0

We can go by the common definition posted here

answered Dec 17 2009 at 6:32

mathfool
1

You should edit the question to include this link there instead of posting it in an answer. – Alberto García-Raboso Dec 17 2009 at 20:05

How does the Dirichlet process work?

Remember to vote up questions/answers you find interesting or helpful (requires 15 reputation points)

3 Answers

You can accept an answer to one of your own questions by clicking the check mark next to it. This awards 15 reputation points to the person who answered and 2 reputation points to you.