Gibbs sampling

Here’s a short note on Gibbs sampling, explaining a simple case with two random variables: gibbs.pdf. Source code: gibbs_2x2.py (run this file using Sage). You can browse all the related files here.

My note and source code follows the 2×2 example in this paper (click for a local copy): Casella, George; George, Edward I. (1992). “Explaining the Gibbs sampler”. The American Statistician 46 (3): 167–174.

Wolfram fail

So Wolfram want people to produce free content for their demonstrations project. Fine. But sending spam where they misquote the title of my paper isn’t the best way to get things done. And if anyone is interested, the full Python source code is already available here. I used SymPy for some symbolic manipulations. It’s all free and open source.

from: Wolfram Research
to: carlo.hamalainen at gmail.com
date: Mon, Nov 9, 2009 at 6:59 PM
subject: Your publication in Arxiv

Dear Carlo Hamalainen,

Your article, “An enumeration of equilateral triangle”, caught the attention of one of my colleagues, who thought that it could be developed into an interesting Demonstration to add to the Wolfram Demonstrations Project. http://demonstrations.wolfram.com/e/2/

The Demonstrations Project, launched alongside Mathematica 6 in May 2007, is a collection of over 5,000 interactive Demonstrations that cover myriad subjects, interests, and skill levels. The Demonstrations are free to download and manipulate thanks to Mathematica Player, which is also free, available at: http://www.wolfram.com/e/2/player

Building a Demonstration is a simple and straightforward process. If you have little or no experience with Mathematica, you may want to attend one of our free seminars. In our S14 seminar, “Creating Demonstrations,” members of the Demonstrations team guide you step-by-step through the authoring process. Learn more about this seminar at: http://www.wolfram.com/e/2/s14.html

Your published Demonstrations will appear on the Wolfram Demonstrations Project website, which averages over 50,000 hits a week.

We welcome any questions you might have, and look forward to seeing a Demonstration submission from you soon.

Best regards,

Marty McKee
Wolfram Research, Inc.
http://www.wolfram.com

Archived Comments

Date: 2009-11-18 12:58:18 UTC

Author: Nadiah

lol “caught the attention of one of my colleagues” … one of my colleagues who happens to be a data-mining algorithm that’s so poorly written it can’t even extract the title of an arxiv paper correctly!

“An enumeration of equilateral triangle”. I bet that was a riveting article:-


Introduction: In this paper we enumerate an equilateral triangle.

Results: There’s only one of them!

Conclusion: There is one equilateral triangle.

Dr Carlo, will your group be applying your novel technique to other related open problems, like enumerating equilateral rectangles or maybe even pentagons? Perhaps you should submit the algorithm to the Wolfram Demonstrations Project:-


def main():
print ‘1’
if __name__ == “__main__”:
main()

Spamin’ Wolfram: A New Kind of Incompetence