Archive for mixture estimation

Newer Entries »

plagiarism exposed!

Posted in R, Statistics, University life with tags , , , , , on June 15, 2010 by xi'an

Last morn, I had the surprise of receiving the following email:

This is to inform you that the following abstract has been submitted to the 3rd International Conference of the ERCIM WG on COMPUTING & STATISTICS (ERCIM’10)

Ab#: 114
Title: Goodness of Fit Via Mixtures of Beta distributions
Keywords: nonparametric estimation, posterior conditional predictive p-value.
Abstract: We consider a Bayesian approach to goodness of fit, that is, to the problem of testing whether or not a given parametric model is compatible with the data at hand . Since we are concerned with a goodness of fit problem, it is more of interest to consider a functional distance to the tested model d(F;F) as the basis of our test, rather than the corresponding Bayes factor, since the later puts more emphasis on the parameters. It is both of high interest and of strong difficulty to come up with a satisfactory notion of a Bayesian test for goodness ofit to a distribution or to a family of distributions.

The abstract is a plagiarism of your work.

I am informing you of about this in case the author has tried to plagiarize the whole paper. The same author has submitted a second abstract plagiarizing another paper.  The author uses bogus affiliations and I cannot trace his institution in case he has one.

It is somehow comforting to see that such a gross example of plagiarism can get detected, despite the fact that our paper never got published. Although I am sure there must be conferences that do not apply any filter on the submission…

This paper with Judith Rousseau was once submitted to Series B, but I could not come to complete the requested revision for programming motives, the task of modifying the several thousand lines of C code driving the beta mixture estimation filling me with paralysing dread! This is actually the time when I stopped programming in C (the fact that I ever really programmed in C is actually debatable!). This is unfortunate as the spirit of the paper was quite nice, using an idea borrowed from Verdinelli and Wasserman to build a genuine Bayesian goodness of fit test… I do not think there is much to salvage at this later stage, given the explosion of Bayesian non-parametrics.

8 Comments »

Typo in Bayesian Core [again]

Posted in Books, R, Statistics with tags , , , , , on May 16, 2010 by xi'an

Reza Seirafi from Virginia Tech sent me the following email about Bayesian Core, which alas is pointing out a real typo in the reversible jump acceptance probability for the mixture model:

With respect to the expression provided on page 178 for the acceptance probability of the split move, I was wondering if the omission of the density of the auxiliary parameters u1, u2, and u3 — specially u2, since its density is not necessarily equal to 1 contrary to the other two — is a case of typo.

This is truly a typo, the acceptance probability at the bottom of page 178 should be

\min\left( \dfrac{\widetilde\pi_{(k+1)k}}{\widetilde\pi_{k(k+1)} }\,\dfrac{\varrho(k+1)}{\varrho(k)}\,\dfrac{\pi_{k+1}(\theta_{k+1})\ell_{k+1}(\theta_{k+1})}{pi(u_2)\pi_{k}(\theta_{k})\ell_{k}(\theta_{k})}\,\dfrac{p_{jk}}{(1-u_1)^2}\,\sigma_{jk}^2,1\right)

since u2 is indeed distributed from a non-uniform density. The physical reason for this (unacceptable!) typo is that I cut-and-pasted the LaTeX code from Monte Carlo Statistical Methods: (page 439) where u2 is a uniform variate! (Incidentally, there is a minor typo a few lines above: when defining

\mu_{(j+1)(k+1)} = \mu_{jk} - \dfrac{p_{j(k+1)}u_2}{p_{jk}-p_{j)(k+1)}},

it should be

\mu_{(j+1)(k+1)} = \mu_{jk} - \dfrac{p_{j(k+1)}u_2}{p_{jk}-p_{j(k+1)}},

another side casualty of cut-and-paste!) Obviously, this could also affect the associated R code but I checked it and found the line

jacob=(kprop-2)*log(1-propp[kprop]) - singleprior(propmix,kprop)

which I think is correcting for the normal proposal.

This typo will obviously be corrected in the next printing of Bayesian Core as well as in the new edition we plan to write in July. It also illustrates the folk theorem “One can never write the reversible jump acceptance probability right” that make me avoid using reversible jump each time I am facing a small enough collection of models to consider the exhaustive list of those models.

Leave a comment »

Mixture estimation and applications

Posted in Statistics, Travel with tags , , , , , , , on March 2, 2010 by xi'an

The meeting on mixtures we organised with Kerrie Mengersen and Mike Titterington at ICMS, Edinburgh, will start tomorrow. While the meeting per se is by invitation only, there will be two public lectures (by definition open to anyone!, so feel free to attend if you are in Edinburgh):

Public Lecture by Kerrie Mengersen (Queensland University of Technology) Doors at 15 South College Street, Edinburgh will open at 17.30 on Wednesday 3 March for a Public Lecture entitled, “Where are they and what do they look like?  Discovering patterns in data using statistical mixture models”  [abstract] The lecture will commence at 18.00 prompt.  Please note that latecomers will not be permitted into the lecture theatre.

Distinguished Lecture by Michael Jordan (University of California, Berkeley) This lecture will be held in Room G07, the Informatics Forum, 10 Crichton Street, Edinburgh on Thursday 4 March at 16.30as part of the Informatics Distinguished Lecture series. It is entitled, “Applied Bayesian nonparametrics” [abstract]

8 Comments »

Mixture estimation and applications

Posted in Statistics, University life with tags , , , , on November 18, 2009 by xi'an

As posted earlier, there will be a workshop on mixture estimation at ICMS, in Edinburgh, in March 2010. It is organised by Kerrie Mengersen and Mike Titterington, and myself. The list of speakers so far is available on the workshop webpage and includes a sufficient array of sensitivities to make the starting roundtable on label switching a lively moment!

While attendance at this workshop is by invitation only and the total attendance will be strictly limited, as the  workshop will be held at 15 South College Street, Edinburgh, with all lectures being given in the Newhaven Lecture Theatre, if you wish to be considered for an invitation please contact within the coming days Professor Mike Titterington at Glasgow University.

4 Comments »

Newer Entries »